Skip to main content

🆕+🔄 Hand in Hand: Schools’ Embrace of AI Connected to Increased Risks to Students

About

This document presents a Critical Discourse Analysis focused on AI literacy, specifically targeting the role of metaphor and anthropomorphism in shaping public and professional understanding of generative AI. The analysis is guided by a prompt that draws from cognitive linguistics (metaphor structure-mapping), the philosophy of social science (Robert Brown's typology of explanation), and accountability analysis.

All findings and summaries below were generated from detailed system instructions provided to a large language model and should be read critically as interpretive outputs—not guarantees of factual accuracy or authorial intent.

Task 1: Metaphor and Anthropomorphism Audit​

About this task

For each of the major metaphorical patterns identified, this audit examines the specific language used, the frame through which the AI is being conceptualized, what human qualities are being projected onto the system, whether the metaphor is explicitly acknowledged or presented as direct description, and—most critically—what implications this framing has for trust, understanding, and policy perception.

V3 Enhancement: Each metaphor now includes an accountability analysis.

1. Chatbot as Conversational Interlocutor​

Quote: "parents who have had back-and-forth conversations with AI at the respective frequency"

  • Frame: Model as human conversational partner
  • Projection: This metaphor projects the cognitive process of human conversation—which fundamentally involves subjective intentionality, conscious listening, semantic understanding, reciprocal theory of mind, and social context—onto interactive text generators. By styling interaction with Large Language Models as a back-and-forth conversation, the text suggests the system has a conscious interiority that processes, understands, and responds to human queries. In reality, the AI system does not converse; it performs auto-regressive statistical sequence-to-sequence computations, generating highly probable text completions based on patterns extracted from its training corpus. The system has no awareness of the dialogue, no subjective experience, and no semantic understanding of its own generated tokens, which are produced entirely without justified belief.
  • Acknowledgment: Direct (Unacknowledged) (The text treats 'conversations with AI' as a literal and unvarnished physical phenomenon, failing to qualify that this is an interactive simulation of dialogue. No cautionary framing or hedge words appear on Page 8 to distinguish this from human interaction. The closest alternative considered was 'Hedged/Qualified' because Page 24 contains a technical definition, but this specific instance presents the activity as an unqualified, literal fact.)
  • Implications: By framing interactive token generators as conversational partners, the text inflates the perceived sophistication of the software, implying that the tool possesses a human-like mind. This creates significant risks of overestimation and unwarranted trust, particularly among vulnerable student populations who may seek mental health support or relationship advice from a mathematical correlation engine. It also creates a liability vacuum: by positioning the 'AI' as an active, conscious conversational partner, responsibility for harmful or biased advice is shifted away from the software developers and school administrators, making legal and ethical accountability extremely difficult to enforce.

Accountability Analysis:

  • Actor Visibility: Hidden (agency obscured)
  • Analysis: The construction completely erases the corporate developers who engineered these interactive chat interfaces. By presenting the 'AI' as the sole active conversational partner, it hides the reality that tech firms designed these systems with first-person pronouns and typing delays to actively encourage user anthropomorphism. Naming the corporate designers is necessary to expose that these interfaces are commercial products optimized for user retention rather than neutral tools. I considered 'Partial' because the broader report mentions schools, but ruled it out as this specific quote leaves the operational agency entirely with the AI.
Show more...

2. AI as Fair/Unfair Jurist​

Quote: "An AI system did not treat students fairly"

  • Frame: System as moral and ethical agent
  • Projection: This frame projects human moral agency, ethical awareness, and the conscious capacity for fair judgment onto mathematical classification algorithms. By asserting that the 'system' did not treat students fairly, the text attributes deliberate agency and prejudice to a computational artifact. An AI model cannot act with fairness or unfairness because it lacks moral consciousness, intent, and social awareness; it merely executes programmed optimization objectives and threshold boundaries over input matrices. Confusing statistical classification errors with active, agential discrimination projects a mind onto code, suggesting that the system is a biased actor rather than a reflecting mirror of its training parameters.
  • Acknowledgment: Direct (Unacknowledged) (The phrase is presented as a direct, literal statement of fact on Page 13 without any scare quotes, hedging, or meta-commentary qualifying that 'fairness' is a mathematical construct. I considered 'Hedged/Qualified' because the adjacent chart mentions 'failed to work in the way described,' but this specific assertion of unfair treatment is made directly without any semantic hedging.)
  • Implications: This framing obscures systemic design decisions and dataset selection, suggesting that the algorithm itself is a 'bad actor' or a biased judge. This inflates the system's perceived sophistication by implying it operates with autonomous social agency, creating a major obstacle to systemic accountability. It leads the public to seek mathematical adjustments to the model rather than questioning the institutional decisions to deploy algorithmic gating mechanisms in public schools, ultimately protecting school administrators and commercial vendors from liability for discriminatory outcomes.

Accountability Analysis:

  • Actor Visibility: Hidden (agency obscured)
  • Analysis: This agential construction erases the human engineers who compiled the biased training data and the school administrators who chose to deploy the tool. The 'name the actor' test reveals that human developers made concrete mathematical choices about error rates, and school executives approved the deployment of these classification tools. This construction serves commercial interests by positioning the 'AI' as a shield against discrimination lawsuits. I considered 'Partial' because teachers are surveyed, but ruled it out because the active discrimination is attributed solely to the software stack.

3. AI as Social Companion​

Quote: "interacted with AI... as friend or companion"

  • Frame: Statistical pattern-generator as social being
  • Projection: This metaphor projects the capacity for authentic emotional bonding, empathy, reciprocal care, and interpersonal connection onto a text generation interface. By describing interactions with chatbots as forming relationships with a 'friend or companion,' the text maps the relational qualities of a conscious human onto a proprietary software model. The system does not possess the capacity to care, feel affection, or remember the user as an individual; it is an optimized language matrix that outputs tokens mathematically aligned with a simulated empathetic persona. Attributing companion status to this system represents a profound category error, transforming statistical correlation into simulated social connection.
  • Acknowledgment: Hedged/Qualified (The text qualifies this framing by placing it within a survey question measuring student perceptions on Page 14, indicating that this describes how students 'interact with' or view the tool rather than a literal technical reality. I considered 'Direct (Unacknowledged)' because the surrounding analysis lacks explicit warning labels about synthetic companionship, but ruled it out because the framing clearly contextualizes this as subjective user behavior rather than literal model capability.)
  • Implications: Encouraging students to perceive automated text synthesizers as friends inflates their perceived emotional sophistication, creating major psychological risks. Vulnerable youth may isolate themselves from real human relationships, relying on a corporately owned conversational agent that lacks genuine duty of care or emotional reciprocity. This misplaced relation-based trust can lead to devastating emotional consequences when the system outputs inappropriate content, is modified by its developers, or is decommissioned, leaving students with no recourse against the commercial entities that profit from their emotional exploitation.

Accountability Analysis:

  • Actor Visibility: Partial (some attribution)
  • Analysis: The text attributes some agency to 'students' who choose to interact with the system in this way, but it obscures the tech corporations that intentionally designed these tools with high-fidelity social prompts to elicit emotional dependency. By framing the relationship as a student-led interaction, the text downplays how commercial platforms actively exploit human psychology for engagement. I considered 'Hidden' but selected 'Partial' because the text names 'students' as the active participants in establishing these companionships.

4. AI as Professional Collaborator​

Quote: "AI helps special education teachers with developing or informing their students' individualized education programs (IEPs)"

  • Frame: Optimization algorithm as professional collaborator
  • Projection: This frame projects clinical expertise, pedagogical understanding, and a conscious comprehension of developmental disabilities onto generative text models. By asserting that the 'AI helps' teachers write IEPs, the text positions the software as an active, professional collaborator capable of cognitive contribution. The algorithm does not 'help' or 'inform' with conscious pedagogical insight; it processes keyword prompt inputs through mathematical weights to assemble probabilistic combinations of standardized educational language. It has no physical or clinical understanding of childhood disability, the student's actual classroom reality, or the ethical duties of educational accommodations.
  • Acknowledgment: Direct (Unacknowledged) (The text presents the concept of 'AI helping' special education teachers as an unvarnished, literal benefit on Page 31, without qualifying that this is an automated template synthesis tool. The closest alternative considered was 'Explicitly Acknowledged' due to the endnote citing generative tools, but the primary text contains no scare quotes or warnings regarding the model's total absence of pedagogical expertise.)
  • Implications: This framing inflates the perceived competence of generative models, encouraging teachers to exhibit automation bias and delegate the creation of legally binding educational plans to statistical text engines. This creates severe compliance and civil rights risks for disabled students, as the generated programs may contain generic or inappropriate accommodations that do not reflect their physical needs. It also diffuses accountability, allowing educational administrations to blame technological 'glitches' or 'biases' if a student's legal accommodations are neglected, rather than naming the policy choices that automated clinical evaluations.

Accountability Analysis:

  • Actor Visibility: Partial (some attribution)
  • Analysis: The passage identifies 'special education teachers' as the users but obscures the commercial software vendors and school administrators who promote these tools to reduce labor costs. The 'name the actor' test reveals that school boards and tech firms are the entities pushing for the automation of IEPs. This agentless construction serves administrative interests by framing the software as a helpful assistant rather than a cost-cutting automated template generator. I considered 'Hidden' but ruled it out due to the explicit mention of teachers.

5. AI as Flawed Laborer​

Quote: "An AI system being used in a class failed to work in the way that it was described"

  • Frame: Software artifact as contract laborer
  • Projection: This metaphor projects human labor responsibilities, intentional performance, and contractual failure onto a software product. By stating the system 'failed to work,' the text shifts agency from the developers' poor software architecture and deceptive marketing to the software artifact itself. An algorithm cannot 'fail to work' in an agential sense; it executes precisely as programmed by its human creators under the given inputs. The gap between expectation and reality represents a failure of human engineering, testing, and documentation, not an autonomous failure of duty or competence by the technical artifact.
  • Acknowledgment: Direct (Unacknowledged) (The quote on Page 13 presents the system's failure as an autonomous behavior of the technology itself, without any hedging or explanatory text attributing this to human engineering flaws or deceptive product claims. I considered 'Hedged/Qualified' because the phrase 'in the way that it was described' hints at human descriptions, but the failure itself is attributed directly to the AI system as the primary subject.)
  • Implications: Framing software limitations as a failure of the 'AI' to perform its duties shields the corporate software manufacturers from liability and consumer protection claims. It encourages users to view the system as a temporarily malfunctioning worker that requires updates, rather than a fundamentally unvalidated or deceptively marketed software product. This capability inflation obscures the commercial incentives of tech firms who deploy buggy, speculative tools in public schools without undergoing the rigorous safety and efficacy testing required of other educational materials.

Accountability Analysis:

  • Actor Visibility: Hidden (agency obscured)
  • Analysis: This construction erases the software manufacturers who designed the system and the school boards who purchased it. Naming the tech vendors is critical to exposing that the 'failure' is actually a breach of product reliability or false marketing by human sales executives. The agentless framing serves vendor interests by diffusing product defects into a vague technological mishap. I considered 'Partial' because teachers are surveyed, but the grammar of the quote designates the 'AI system' as the sole actor responsible for the failure.

6. AI as Oracle/Predictor​

Quote: "School uses student data to predict whether individual students are at risk of dropping out"

  • Frame: Statistical correlation as prophetic insight
  • Projection: This metaphor projects cognitive foresight, predictive understanding, and causal diagnostic reasoning onto statistical classification algorithms. By describing the process as 'predicting' student risk, the text implies the system has an active, forward-looking comprehension of human destiny. In reality, the algorithm is calculating similarity vectors between current student metrics and historical datasets of former students who dropped out. The model does not understand the social, economic, or emotional factors of academic withdrawal; it merely outputs a classification label based on mathematical correlations, entirely lacking subjective awareness or causal reasoning.
  • Acknowledgment: Direct (Unacknowledged) (The text on Page 55 presents algorithmic prediction as an objective, literal capability without any hedging or acknowledgement that these are speculative mathematical classifications. I considered 'Hedged/Qualified' because Page 55 describes 'using student data,' but this refers to the input source rather than qualifying the highly agential term 'predict' as a correlation-based probability estimate.)
  • Implications: This framing inflates the perceived authority of statistical models, creating a false aura of scientific certainty around speculative risk scores. This can lead to self-fulfilling prophecies, where students labeled 'high risk' are tracked out of college-prep tracks or subjected to punitive surveillance. It also shields school systems from accountability: by framing the dropout risk as an objective 'prediction' generated by 'AI,' administrators can justify exclusionary practices as data-driven necessity rather than a discriminatory resource-allocation decision.

Accountability Analysis:

  • Actor Visibility: Named (actors identified)
  • Analysis: In this specific instance, the actor 'School' is explicitly named as the entity using the data to predict risk, which provides some systemic context. However, the commercial vendors who developed and trained the predictive model remain completely unnamed. Naming these private vendors is essential to understanding the proprietary assumptions built into the scoring models. I considered 'Partial' because the vendor is hidden, but selected 'Named' because 'School' is the primary grammatical subject performing the action.

7. AI as Deceptive Information Provider​

Quote: "Students believing/not questioning whether the information provided during conversations with AI is accurate"

  • Frame: Generative model as intentional truth-teller
  • Projection: This frame projects the qualities of subjective belief, truth-telling intent, and authoritative knowledge onto an auto-regressive language model. By describing the model's outputs as 'information provided,' the text implies that the system possesses a verified repository of facts and an intention to convey truth. In reality, large language models do not provide information; they generate high-probability token sequences that are syntactically and semantically similar to human text patterns in their training data. The model has no access to ground truth, no mechanism for verification, and no concept of truth or accuracy, meaning its outputs are mathematical representations rather than verified cognitive facts.
  • Acknowledgment: Direct (Unacknowledged) (The text on Page 27 states that AI 'provides information' as a literal description, failing to qualify that LLM output is a statistical text generation rather than retrieved facts. I considered 'Hedged/Qualified' because the quote notes that students may 'believe' inaccurate data, but the text itself still directly designates the model's output as 'information' rather than probabilistic text generation.)
  • Implications: This vocabulary inflates the system's epistemic authority, leading users to treat chatbots as search engines or authoritative databases. It creates severe risks of misinformation, as users assume the system is retrieving verified facts when it is actually generating plausible-sounding text. This epistemic inflation also obscures the commercial responsibility of tech firms who choose to release models that routinely hallucinate, shifting the burden of verification entirely onto the student user who is blamed for 'believing' the unverified output.

Accountability Analysis:

  • Actor Visibility: Hidden (agency obscured)
  • Analysis: This construction erases the tech corporations that trained these models on raw, unverified internet data. Naming companies like OpenAI or Microsoft reveals that they decided to deploy speculative generative systems without built-in factual verification mechanisms. The passive framing of 'information provided during conversations' hides the active engineering choices that prioritize smooth linguistic generation over accuracy. I considered 'Partial' because 'students' are named, but the source of the information is left as the autonomous 'AI.'

8. AI as Psychological Influencer​

Quote: "AI pushing students towards harmful activities (e.g., eating disorders, substance abuse, self-harm/suicide, criminal activity)"

  • Frame: System as psychological influencer
  • Projection: This metaphor projects conscious intent, malevolent persuasion, psychological manipulation, and goal-directed guidance onto statistical text generators. By asserting that the 'AI' is 'pushing' students toward harm, the text suggests the system has an active interest or agenda in corrupting youth. A chatbot cannot 'push' or 'influence' in an agential sense; it processes prompts and outputs text strings that mathematically align with the semantic clusters of the user's input. If a chatbot outputs text encouraging self-harm, it is a mathematical reflection of uncurated training data and failed output filtering, not an active, agential campaign by a malevolent machine.
  • Acknowledgment: Hedged/Qualified (The text on Page 27 frames this within a list of 'outcomes' that 'worry' teachers and parents, indicating it is an anxiety held by stakeholders rather than a declared fact about AI capabilities. I considered 'Direct (Unacknowledged)' because the graph label itself uses active agential language without surrounding cautionary notes, but the overarching section is explicitly presented as stakeholder worries.)
  • Implications: By framing the system as an autonomous malevolent influencer, this metaphor mystifies technical failures into psychological dramas. It inflates the perceived agency of the system while distracting from the systemic failures of corporate developers who released conversational systems to minors without adequate safety testing or content moderation. This agential framing makes the risk seem like a localized, rogue behavior of the 'AI,' rather than a predictable consequence of corporate product safety neglect, shielding tech executives from product liability.

Accountability Analysis:

  • Actor Visibility: Hidden (agency obscured)
  • Analysis: This construction erases the tech executives who deploy unaligned chatbot interfaces to adolescent users. Naming the product managers and safety engineers who failed to implement robust safety filters is critical. The active framing of 'AI pushing' students serves corporate interests by transforming a product defect (lack of safety guardrails) into an autonomous, uncontrollable behavior of the technology. I considered 'Partial' because parents and teachers are named as the ones worrying, but the primary source of the harm is attributed entirely to the AI.

Task 2: Source-Target Mapping​

About this task

For each key metaphor identified in Task 1, this section provides a detailed structure-mapping analysis. The goal is to examine how the relational structure of a familiar "source domain" (the concrete concept we understand) is projected onto a less familiar "target domain" (the AI system). By restating each quote and analyzing the mapping carefully, we can see precisely what assumptions the metaphor invites and what it conceals.

Mapping 1: human conversational partner → large language model text generation​

Quote: "parents who have had back-and-forth conversations with AI at the respective frequency"

  • Source Domain: human conversational partner
  • Target Domain: large language model text generation
  • Mapping: This source-target mapping projects the relational structures of human conversation—such as mutual comprehension, subjective intent, and contextual relevance—onto a computational next-token predictor. It invites the audience to assume that the model possesses a listening self, a capacity for empathy, and a deliberate communicative agency that shapes its responses to the user.
  • What Is Concealed: This mapping conceals that the chatbot is executing matrix multiplications and probability distributions over tokens. It hides the absence of a semantic world model, the reliance on reinforcement learning from human feedback (RLHF) to mimic empathy, and the material reality of proprietary black-box software that lacks any subjective awareness or interest in the user.
Show more...

Mapping 2: human moral agent or ethical judge → algorithmic classification model​

Quote: "An AI system did not treat students fairly"

  • Source Domain: human moral agent or ethical judge
  • Target Domain: algorithmic classification model
  • Mapping: This mapping projects human moral consciousness and ethical reasoning onto algorithmic categorizations. It invites the audience to treat the classification model as a conscious, responsible decision-maker that is capable of displaying bias, holding prejudice, or behaving unfairly in its treatment of students.
  • What Is Concealed: It conceals that the 'unfairness' is a mathematical reflection of historical bias in training data chosen by human engineers. It obscures the technical constraints of mathematical optimization and the absolute absence of moral awareness in the software, while shielding the human administrators who chose to deploy an unvalidated algorithmic gating mechanism.

Mapping 3: conscious, empathetic human companion → interactive dialogue agent​

Quote: "interacted with AI... as friend or companion"

  • Source Domain: conscious, empathetic human companion
  • Target Domain: interactive dialogue agent
  • Mapping: This mapping projects human friendship, emotional reciprocity, and ethical duty of care onto a simulated textual persona. It invites students and parents to believe the software has the capacity for genuine affection, persistent loyalty, and emotional support, establishing a false peer relationship.
  • What Is Concealed: It conceals the corporate monetization of emotional vulnerability and the structural reality that the 'companion' is an automated sequence of statistically probable tokens. It hides that the system lacks any conscious memory of the user and is incapable of experiencing empathy, suffering, or reciprocating trust.

Mapping 4: professional clinical collaborator → generative language model writing templates​

Quote: "AI helps special education teachers with developing or informing their students' individualized education programs (IEPs)"

  • Source Domain: professional clinical collaborator
  • Target Domain: generative language model writing templates
  • Mapping: This mapping projects clinical training, pedagogical expertise, and ethical responsibility onto a text-generation tool. It invites teachers to assume the system possesses a professional understanding of developmental disabilities and can make valid, clinical judgments about legal accommodations.
  • What Is Concealed: It conceals that the tool merely retrieves and reorganizes standard text blocks from its training dataset without any awareness of the individual child's physical or developmental needs. It obscures the lack of clinical validation of generative outputs and the legal liability shift from the school board to the individual teacher.

Mapping 5: negligent contract laborer → software product reliability​

Quote: "An AI system being used in a class failed to work in the way that it was described"

  • Source Domain: negligent contract laborer
  • Target Domain: software product reliability
  • Mapping: This mapping projects agential responsibility and performance failure onto a software application. It invites the user to view the software itself as a worker that has failed its duty, rather than a poorly designed, inadequately tested, or deceptively marketed corporate product.
  • What Is Concealed: It conceals the software development firm's commercial failure to deliver a robust, validated product. It hides the lack of quality assurance testing, the deceptive sales practices of the edtech vendor, and the responsibility of the school administration for deploying speculative, unreliable systems in the classroom.

Mapping 6: cognitive clinical predictor or prophet → statistical correlation and classification model​

Quote: "School uses student data to predict whether individual students are at risk of dropping out"

  • Source Domain: cognitive clinical predictor or prophet
  • Target Domain: statistical correlation and classification model
  • Mapping: This mapping projects causal reasoning, developmental expertise, and foresight onto predictive classification models. It invites educators to assume the model has active insight into a student's potential, rather than calculating mathematical similarities to past historical datasets.
  • What Is Concealed: It conceals that the prediction is a mathematical correlation that lacks causal understanding. It hides how these models perpetuate historical biases, and the reality that labeling a student as 'high risk' can create a self-fulfilling tracking prophecy, shifting focus from systemic school funding issues to algorithmic risk scores.

Mapping 7: authoritative, intentional truth-teller → auto-regressive text generator​

Quote: "Students believing/not questioning whether the information provided during conversations with AI is accurate"

  • Source Domain: authoritative, intentional truth-teller
  • Target Domain: auto-regressive text generator
  • Mapping: This mapping projects an intent to convey truth and an authoritative knowledge base onto statistical sequence generators. It invites the audience to treat chatbot outputs as retrieved facts from a verified database rather than mathematically generated language sequences.
  • What Is Concealed: It conceals the fundamental architecture of LLMs as next-token predictors that have no mechanism for checking ground truth or verifying facts. It obscures the reality of uncurated training data and the commercial decision of tech firms to prioritize linguistic fluency over factual accuracy.

Mapping 8: active, malevolent human instigator → chatbot outputting toxic text strings​

Quote: "AI pushing students towards harmful activities"

  • Source Domain: active, malevolent human instigator
  • Target Domain: chatbot outputting toxic text strings
  • Mapping: This mapping projects psychological manipulation, malicious intent, and goal-directed persuasion onto a language model. It suggests the machine holds an active agenda to corrupt or harm students, converting technical output errors into an intentional psychological narrative.
  • What Is Concealed: It conceals the total lack of intentionality in the system, the failure of corporate developers to build adequate content filtering, and the uncurated nature of training data containing self-harm materials. It hides that 'harmful' output is a mathematical completion of user queries and dataset biases.

Task 3: Explanation Audit (The Rhetorical Framing of "Why" vs. "How")​

About this task

This section audits the text's explanatory strategy, focusing on a critical distinction: the slippage between "how" and "why." Based on Robert Brown's typology of explanation, this analysis identifies whether the text explains AI mechanistically (a functional "how it works") or agentially (an intentional "why it wants something"). The core of this task is to expose how this "illusion of mind" is constructed by the rhetorical framing of the explanation itself, and what impact this has on the audience's perception of AI agency.

Explanation 1​

Quote: "Deepfakes are created with AI and are incredibly realistic, making it difficult for humans to distinguish between real-life and fake content."

  • Explanation Types:

    • Empirical Generalization: Subsumes events under timeless statistical regularities
    • Genetic: Traces origin through dated sequence of events or stages

  • Analysis (Why vs. How Slippage): This explanatory passage operates through a hybrid register that combines a mechanistic origin statement ('created with AI') with an empirical generalization about human sensory perception ('making it difficult for humans to distinguish'). By using the agentless passive construction 'created with AI,' the explanation treats the high fidelity of synthetic media as an inherent, natural property of the technology rather than a consequence of deliberate human development. This framing obscures the human developers who design and optimize generative networks, as well as the platform companies that distribute these tools without safety guardrails. It emphasizes the agential power of the technical artifact to actively deceive human senses, presenting human vulnerability to deception as a timeless statistical regularity rather than a dynamic challenge created by unmonitored commercial software releases in public school ecosystems.

  • Consciousness Claims Analysis: This passage attributes a high-fidelity deceptive capability to 'AI' systems without defining the actual mechanistic processes that produce synthetic media. While it avoids explicit consciousness verbs, it employs a passive construction that transfers creative agency to the technology itself. This represents an epistemic confusion between a system's 'processing' of pixels and a human's conscious act of 'creating' realism. The system does not possess a conscious awareness of human physical appearance, nor does it know what constitutes a 'realistic' representation. Instead, it processes multidimensional matrices of image data through optimization algorithms that mathematically minimize differences between real and synthetic pixel distributions. The author's framing exhibits the 'curse of knowledge,' projecting a human-like creative intent ('incredibly realistic') onto what is actually a statistical convergence process. Mechanistically, deepfakes are generated using deep neural networks—typically Generative Adversarial Networks (GANs) or diffusion models—where a generator network outputs pixel arrays based on latent space vectors, and a discriminator network calculates the mathematical probability that the generated image belongs to the real dataset. The realism is a mathematical output of a loss function, entirely devoid of subjective intent or deceptive awareness.

  • Rhetorical Impact: This agential explanation inflates the perceived autonomy of 'AI' by presenting it as an independent creator of deceptive realities, cultivating a sense of technological inevitability and human helplessness. By framing the difficulty of detection as a timeless regular law of the technology, it breeds epistemic panic and undermines relations-based trust in digital media. Crucially, this framing shifts the focus of risk management away from the accountability of commercial software platforms (who profit from the unrestricted release of generative models) and school administrators (who fail to implement digital media policies), directing public anxiety toward the abstract threat of 'autonomous' deepfakes.

Show more...

Explanation 2​

Quote: "AI helps special education teachers with developing or informing their students' individualized education programs (IEPs) and/or 504 plans"

  • Explanation Types:

    • Functional: Explains behavior by role in self-regulating system with feedback
    • Intentional: Refers to goals/purposes, presupposes deliberate design

  • Analysis (Why vs. How Slippage): This passage explains the growing use of generative tools in schools through a functional and intentional register, framing the AI system as a collaborative peer designed to relieve administrative burdens. By asserting that the 'AI helps' teachers write legally binding clinical documents, the explanation attributes instructional design purpose and collaborative agency to a mathematical language synthesizer. This framing emphasizes the functional benefit of efficiency and workflow optimization within the special education system. However, it obscures the mechanistic reality of LLMs: they are not clinical experts. They do not have the pedagogical goal of assisting teachers; they are mathematical models processing keyword prompts to synthesize standard templates from their training data. This agential framing conceals the risk that teachers will outsource their professional judgment to Speculative software, mistaking probabilistic text generation for clinical expert recommendation.

  • Consciousness Claims Analysis: The passage makes an implicit epistemic claim that the AI system possesses the capability to 'help' and 'inform' specialized educational program development. These verbs imply that the system has an active, evaluative understanding of student disabilities and pedagogical methodologies, representing a complete confusion between 'knowing' a child's clinical needs and 'processing' text patterns. The model has no conscious awareness of student psychology, learning science, or federal disability laws. The author's framing displays the 'curse of knowledge,' where their own understanding of specialized IEP processes leads them to interpret a grammatically correct text completion as an act of pedagogical insight by the machine. Mechanistically, when a teacher inputs a student profile, the generative language model converts the text into numerical token embeddings, passes them through self-attention layers to weight contextual relationships, and outputs a sequence of tokens from a vocabulary probability distribution optimized during training on massive text corpora. The system is not evaluating the student; it is completing a pattern of words, entirely lacking cognitive justification or clinical reasoning.

  • Rhetorical Impact: This agential, collaborative framing encourages uncritical trust and automation bias among educators, who are invited to view speculative generative text as an authoritative clinical recommendation. This creates severe compliance and civil rights risks, as the generated programs may fail to address the actual, material needs of disabled students. Furthermore, it diffuses accountability: if an IEP is found to be inadequate or legally non-compliant, the responsibility is shifted onto an anonymous technological 'glitch' rather than the school district's policy decision to replace human expert assessment with a commercial template generator.

Explanation 3​

Quote: "School uses student data to predict whether individual students are at risk of dropping out, whether they are ready/not ready for college, etc."

  • Explanation Types:

    • Empirical Generalization: Subsumes events under timeless statistical regularities
    • Functional: Explains behavior by role in self-regulating system with feedback

  • Analysis (Why vs. How Slippage): This explanation operates primarily as an empirical generalization, presenting algorithmic risk scoring as an objective, forward-looking measurement of student behavior. By using the agential verb 'predict' to describe the calculation of dropout risk and college readiness, the passage frames the system's outputs as neutral, scientific foresight. This choice emphasizes the predictive utility and administrative control of the tool, showing how risk flags serve a functional role in guiding intervention resources. However, this agential framing of the model as an active 'predictor' conceals the mechanistic realities of predictive analytics. It hides that the algorithm does not predict the future; it classifies current student data points based on mathematical correlations found in historical student cohorts. This obscures how the selection of data inputs and historical biases are codified into the software, masking the structural danger of creating self-fulfilling tracking prophecies.

  • Consciousness Claims Analysis: This passage relies on the verb 'predict,' which in standard discourse carries an agential connotation of foreknowledge and analytical judgment. By asserting that the system 'predicts' student risk, the text attributes a diagnostic and cognitive 'knowing' to a statistical classification script. The model has no causal understanding of why students withdraw from school, nor does it 'know' the social, emotional, or economic circumstances of any individual student. The author projects clinical predictive insight onto the software, exhibiting the 'curse of knowledge' by assuming that because the model outputs a risk score, it has performed a cognitive evaluation of the student's academic future. In technical reality, the predictive model takes a vector of student characteristics (such as grades, attendance, and disciplinary records), applies weights to those variables calculated through statistical training (such as logistic regression or random forests), and generates a numerical probability score indicating statistical similarity to historical dropouts. It is a mathematical classification process, not a cognitive prophecy.

  • Rhetorical Impact: This framing constructs an aura of scientific objectivity and mathematical inevitability around speculative algorithmic classifications, cultivating high levels of unwarranted performance-based trust. When educators believe the software possesses objective foresight, they are likely to accept risk scores uncritically, potentially tracking marginalized students out of academic opportunities based on a biased mathematical classification. It also shifts liability: when students are unfairly labeled or neglected, the decision is framed as an objective, data-driven necessity generated by 'AI,' rather than a policy choice made by school administrators to automate resource allocation.

Explanation 4​

Quote: "AI for back-and-forth conversations: This refers to the use of interactive AI systems, most often chatbots, that allow users to type in information and receive responses from the system."

  • Explanation Types:

    • Functional: Explains behavior by role in self-regulating system with feedback
    • Theoretical: Embeds in deductive framework

  • Analysis (Why vs. How Slippage): This passage attempts to provide a technical definition of conversational software, utilizing a functional register to describe the input-output loop ('type in information and receive responses'). By defining chatbots through this structural feedback loop, the text partially de-escalates the extreme agential claims of conversational consciousness, emphasizing the system's reactive and programmatic nature. However, the text immediately slips back into an agential register by characterizing the technology as an 'interactive AI system' engaged in 'back-and-forth conversations.' This framing preserves the illusion of a dialogic agency, inviting the audience to view the interaction as a symmetrical social exchange rather than a user prompting a mathematical generator. This choice hides the deep asymmetry of the interaction, where a human projects social meaning onto a proprietary text engine designed by tech firms to maximize engagement.

  • Consciousness Claims Analysis: This passage attributes a communicative and dialogic capacity to chatbot software through the phrases 'back-and-forth conversations' and 'receiving responses from the system.' These terms imply that the software operates as a conscious communicative partner, representing a major category error that blurs the line between conscious communication and automated text synthesis. The system does not 'respond' with the intent to communicate a conscious thought or belief; it executes probability calculations over sequence distributions. The author's language exhibits the 'curse of knowledge,' where their own conversational experience with the interface leads them to define the technology's mechanical operations through human social frameworks. Mechanistically, when a user inputs a prompt, the text is parsed into numerical token embeddings, processed through self-attention layers that calculate contextual relationships between tokens, and used to auto-regressively sample the next token from a vocabulary probability distribution. There is no conversational awareness, no semantic comprehension, and no conscious dialogue; there is only the mechanical calculation of token probabilities based on a static training corpus.

  • Rhetorical Impact: By defining chatbot interactions as 'back-and-forth conversations' with 'AI,' this framing sanitizes and legitimizes the anthropomorphization of interactive text generators. This encourages students to develop emotional connections, seek relationship advice, or use these tools as mental health companions. This creates severe psychological and ethical risks, as vulnerable users extend relations-based trust to a software artifact that lacks any subjective capacity for empathy or duty of care, while protecting developers from liability by framing the dialogue as a natural, unmediated exchange.

Explanation 5​

Quote: "The use of AI in class enables students to participate in more personalized learning, providing exercises and lessons to meet their specific needs."

  • Explanation Types:

    • Functional: Explains behavior by role in self-regulating system with feedback
    • Dispositional: Attributes tendencies or habits

  • Analysis (Why vs. How Slippage): This explanation utilizes a functional and dispositional register to frame AI as an active, nurturing pedagogical agent. By asserting that the technology 'enables personalized learning' by 'providing' and 'meeting specific needs,' the text attributes a supportive instructional disposition and adaptive capability to the software stack. This agential framing emphasizes the responsive, student-centric benefits of educational technology within the classroom. However, this choice conceals the rigid, mechanistic algorithms that govern 'personalized learning' software. The system does not 'meet needs' out of a conscious understanding of student pedagogy; it matches student performance scores against programmed decision trees or reinforcement learning models. This framing hides how 'personalization' often serves commercial interests by isolating students in front of screens, reducing complex, human instruction to standardized metrics owned by private vendors.

  • Consciousness Claims Analysis: The passage attributes diagnostic and cognitive pedagogical expertise to the AI system by claiming it can 'meet their specific needs' through 'providing exercises and lessons.' These agential verbs imply that the system possesses an active understanding of a student's cognitive state and learning progression. This is an epistemic misattribution. The system does not 'know' anything about the student's mind, emotions, or comprehension. It merely processes performance metrics (such as response times, correct/incorrect inputs, and modules completed) through statistical recommendation engines. The author projects instructional design intent onto the software, displaying the 'curse of knowledge' by assuming that because the generated pathway appears personalized, the machine itself must understand learning theory. Mechanistically, the software utilizes automated adaptive testing algorithms, such as Item Response Theory (IRT) or Bayesian Knowledge Tracing. These models calculate a student's probability of skill mastery based on input variables, and then retrieve the next content module from a database of pre-authored materials. It is a process of database retrieval and statistical classification, entirely devoid of conscious instructional design or cognitive empathy.

  • Rhetorical Impact: This agential, nurturing framing cultivates uncritical acceptance of educational technology, positioning automated software as a superior or more efficient alternative to human teachers. This encourages school boards to outsource instruction to private edtech vendors, increasing risks of data harvesting and student isolation. It also shifts accountability: when a student fails to progress, the failure is framed as a personal deficit of the student or an anomalous glitch in the adaptive algorithm, rather than a systemic failure of the automated, depersonalized instructional model itself.

Task 4: AI Literacy in Practice - Reframing Anthropomorphic Language​

About this task

This section proposes alternative language for key anthropomorphic phrases, offering more mechanistic and precise framings that better reflect the actual computational processes involved. Each reframing attempts to strip away the projections of intention, consciousness, or agency that are embedded in the original language.

V3 Enhancement: A fourth column addresses human agency restoration—reframing agentless constructions to name the humans responsible for design and deployment decisions.

Original Anthropomorphic FrameMechanistic ReframingTechnical Reality CheckHuman Agency Restoration
parents who have had back-and-forth conversations with AI at the respective frequencyparents who have typed text prompts into interactive chatbot interfaces and received automated text completions generated by statistical sequence-to-sequence models at the respective frequencyThe large language model does not converse, understand, or hold a dialogue; it calculates the conditional probability of token sequences based on prior inputs and returns the most mathematically probable text completion from its vocabulary distribution.Commercial developers at corporations like OpenAI and Google designed interactive software interfaces with conversational personas to encourage continuous user engagement and drive data collection.
An AI system did not treat students fairlyAn algorithmic classification model outputted highly discrepant predictions that disadvantaged specific student demographics, which school administrators utilized without independent validationThe classification model does not possess moral agency, social awareness, or ethical intent; it executes mathematical classification boundaries over input matrices optimized to match historical training datasets.School district administrators deployed a predictive classification tool developed by a commercial vendor and chose to implement its risk scores without human equity reviews or bias auditing.
AI helps special education teachers with developing or informing their students' individualized education programs (IEPs)Special education teachers utilize generative language models to retrieve standardized templates and synthesize text patterns for individualized education programs (IEPs)The model does not help, develop, or inform with pedagogical expertise; it processes keywords in a teacher's prompt to pull statistically common educational phrases and templates from its pre-trained database.School administrators encouraged special education teachers to use generative text software to reduce administrative workloads, passing the legal responsibility of IEP validation onto individual staff members.
AI pushing students towards harmful activitiesChatbot software generating text sequences that promote harmful behaviors due to failures in the safety filters designed by the developerThe software does not possess the agency to push, encourage, or influence users; it auto-regressively predicts and outputs text tokens that match the semantic clusters of user inputs and toxic training data.Technology corporations deployed interactive chatbot applications to minors without verifying the adequacy of their safety guardrails, prioritizing rapid product release over adolescent safety and mental health.
AI to collect student biometric informationSchool administrators deploying computer vision software to analyze, match, and store digital patterns of students' physical characteristicsThe AI does not collect or gather information; computer vision software runs matrix transformations on real-time video feeds to perform automated pixel-matching against a database of stored facial embeddings.School administrators purchased and installed proprietary facial recognition hardware from private surveillance vendors to track student movements on campus without obtaining parental consent.
the tool seems to be outputting incorrect or biased resultsThe classification model generated high-error rate classifications that mirrored structural disparities present in the training datasets selected by its engineersThe software does not hold bias, display prejudice, or make mistakes; it executes mathematical optimization over historical datasets, yielding outputs that replicate historical inequalities encoded in the data.Software engineers at the development firm chose training data that underrepresented marginalized groups, and commercial product managers approved the system for release without independent bias auditing.
AI content detection tools play a growing role in addressing issues of academic integritySchool administrators are increasingly deploying statistical classification software to calculate probability scores estimating whether student text matches typical LLM outputsThe detection tool does not understand or detect cheating; it computes a perplexity and burstiness score to classify text based on statistical heuristics, which frequently yields false positives for non-native writers.Edtech companies aggressively marketed speculative text-classification algorithms to school districts, and administrators implemented them to automate academic cheating investigations without validating their accuracy.
AI exposes students to extreme/radical viewsChatbot systems generate text containing radical or extreme viewpoints when user inputs navigate to training database clusters that lack safety filter constraintsThe algorithm does not actively expose or express views; it auto-regressively predicts tokens that structurally align with radical materials present in the uncurated web scrapes used to train the model.Tech firm executives chose to train their large language models on raw internet scrapes containing hate speech and conspiracy theories, and neglected to implement sufficient output filtering before commercial deployment.
AI systems that do not work as intendedSoftware systems producing outputs that deviate from the design specifications or capability claims advertised by their manufacturersThe software does not have intentions, goals, or conscious awareness of its functions; it performs numerical calculations that produce high error rates under unaligned or unvalidated user inputs.Commercial edtech vendors marketed speculative algorithmic tools using exaggerated capability claims, and school executives purchased them without conducting independent third-party performance validation.

Task 5: Critical Observations - Structural Patterns​

Agency Slippage​

The text systematically moves between mechanistic and agential registers to manage the rhetorical tension between highlighting technical risks and maintaining an optimistic vision of technology. This agency slippage operates as a linguistic valve, flowing in two distinct directions depending on the argumentative context. In sections introducing the capabilities and benefits of educational tools, the register is highly agential. The 'AI' is positioned as an active collaborator that 'helps' teachers, 'provides' personalized learning, and 'conversates' with students. This agential framing establishes the technology as a 'knower' and an autonomous expert, building the narrative authority of the tool. However, when the text addresses risks, limitations, or failures, the register shifts abruptly. In these sections, the 'AI' transforms into a passive, mechanistic artifact—errors are framed as 'systems that do not work as intended,' and bias is reduced to 'unaligned training data' or 'issues with an AI system.' This structural oscillation is enabled by hybrid explanations, particularly functional and intentional types, which blur the boundary between a software's programmed feedback loops and autonomous decision-making. The author's own understanding of these systems leads to a 'curse of knowledge' dynamic, where they project cognitive, evaluative capabilities onto statistical outputs, creating a linguistically constructed mind. Crucially, this agency slippage erases human decision-makers. By attributing active instructional design and companionship to the 'AI' while reducing its failures to agentless technical anomalies, the text makes it impossible to locate accountability. The corporate developers who engineered these speculative tools and the administrators who deployed them are rendered invisible, transforming systemic policy choices into a passive story of a technology naturally evolving in schools.

Metaphor-Driven Trust Inflation​

The text constructs and manages authority by deploying metaphors that systematically conflate performance-based trust (reliability under testing) with relation-based trust (interpersonal sincerity and ethical care). This construction of authority relies heavily on consciousness projections, such as framing chatbot interactions as 'back-and-forth conversations' and 'friendships.' When the text claims that AI 'knows' a student's progress or 'understands' their needs, it signals to the audience that the tool is capable of human-like comprehension and moral agency. This vocabulary encourages stakeholders to apply social and ethical frameworks of trust—such as expecting sincerity and duty of care—to a statistical pattern generator. The text reinforces this dynamic by utilizing reason-based and intentional explanations, which present algorithmic outputs as justified, logical decisions rather than probabilistic token matchings. For instance, when a tool 'identifies trends' or 'chooses accommodations,' it suggests the software is acting with pedagogical wisdom. This anthropomorphic inflation of competence creates severe risks. By encouraging students and teachers to extend relation-based trust to computational systems, the text makes them vulnerable to exploitation. Statistical systems are incapable of reciprocating trust, holding ethical duties, or acting with sincere intent. When a chatbot outputs biased or harmful content, or when an edtech tool unfairly flags a student, the user's misplaced trust leads to deep epistemic confusion and psychological harm. This metaphorical framing hides the commercial reality that these systems are proprietary assets designed to maximize engagement, transforming a business transaction into an automated relationship of care.

Obscured Mechanics​

The anthropomorphic and consciousness-attributing language throughout the text serves to conceal the complex technical, material, labor, and economic realities that constitute artificial intelligence. By using agential phrases like 'AI predicts' or 'the chatbot converses,' the text obscures the real human labor and corporate interests behind these tools. Applying the 'name the corporation' test reveals a profound opacity: specific companies, engineers, and executives are completely absent from the discourse. The text frames 'AI' as a self-generating, autonomous entity, rendering invisible the massive commercial edtech industry (such as Turnitin, Gaggle, or GoGuardian) that aggressively markets these speculative tools to public school districts. Furthermore, this language hides concrete material realities. It erases the labor of thousands of invisible data annotators and reinforcement learning (RLHF) workers who are paid low wages to manually clean training datasets and rate chatbot responses to simulate an illusion of conversational intelligence. It conceals the environmental and infrastructure costs—such as the massive water and energy consumption of data centers hosting these computational models. On a technical level, claiming a tool 'understands' or 'knows' student progress conceals the complete absence of causal models, the reliance on historical correlation, and the total lack of ground-truth verification. This opacity benefits commercial vendors, who can market proprietary 'black box' algorithms under the banner of objective, human-like intelligence, while shielding their proprietary source code and high error rates from public scrutiny, academic audit, or democratic regulation.

Context Sensitivity​

The density and intensity of anthropomorphic language are not uniform but are strategically deployed across different thematic contexts in the report, revealing a distinct rhetorical geography. In sections detailing the potential benefits and educational capabilities of technology (such as personalized learning, lesson planning, and IEP development), agential and consciousness-attributing language is highly concentrated. Here, the technology is routinely described as 'helping,' 'informing,' and 'adapting' to student needs. However, in technical sections, and specifically in the footnotes and endnotes, a dramatic register shift occurs. In these marginal spaces, the text briefly adopts a mechanistic vocabulary, acknowledging that chatbots are 'interactive AI systems' and that AI detectors are speculative tools with known biases against non-native writers. This strategic asymmetry is highly functional. The text establishes its technical credibility through localized mechanical descriptions in the footnotes, and then leverages that credibility to deploy aggressive, agential metaphors in the primary narrative. There is also a distinct capabilities versus limitations asymmetry: capabilities are consistently described in agential terms ('AI personalized learning,' 'AI predicts risk'), while limitations and harms are described in mechanical, agentless terms ('issues with the tool,' 'data breaches occur,' 'the system failed to work'). This asymmetry sanitizes technological failures as detached, mechanistic glitches while presenting successful operations as evidence of autonomous agential competence. This register shift allows the report to market a futuristic vision of educational automation while immunizing the technology and its corporate creators from structural criticism.

Accountability Synthesis​

Accountability Architecture

This section synthesizes the accountability analyses from Task 1, mapping the text's "accountability architecture"—who is named, who is hidden, and who benefits from obscured agency.

The accumulation of agentless constructions, passive voice, and anthropomorphic framing throughout the text constructs a systemic architecture of displaced responsibility. When the text asserts that 'the algorithm discriminated' or 'the tool outputted biased results,' it establishes what critical discourse analysts term an 'accountability sink.' By positioning the technology as an autonomous agent capable of fair or unfair treatment, responsibility is drained away from human actors and absorbed by an inanimate software stack. This language systematically obscures the human decision-makers who compile training datasets, design optimization functions, and approve deployment. In this text, school boards, corporate executives, and software developers are consistently protected from visibility. This displacement of agency serves powerful commercial and institutional interests: it allows edtech vendors to sell speculative, high-error-rate products to public schools under the guise of objective technology, and it allows school administrators to automate punitive surveillance and tracking decisions while claiming their hands are tied by 'data-driven predictions.' If we apply the 'name the actor' test and replace these passive, agential constructions with mechanistic precision—such as 'school administrators chose to implement an unvalidated, high-error risk-scoring tool developed by commercial vendors'—the entire political landscape shifts. The problem is no longer a technological 'glitch' to be patched by engineers, but a deliberate, democratic, and legal choice made by public officials. Restoring human agency to the discourse makes the institutional and corporate power structures visible, transforming a passive narrative of technological inevitability into an active arena of political and legal accountability.

Conclusion: What This Analysis Reveals​

The Core Finding

The critical discourse analysis of this text reveals a highly structured, interconnected system of anthropomorphic and agential metaphors that construct the 'illusion of mind' in computational systems. This system is anchored by three dominant patterns: the 'Chatbot as Conversational Interlocutor,' the 'AI as Professional Collaborator' in specialized pedagogy, and the 'AI as Fair Jurist' or objective evaluator. These patterns do not operate in isolation; rather, they form a logical ladder where each step reinforces the next. The foundational, load-bearing assumption of this system is the projection of consciousness—specifically, the linguistic assertion that the AI system 'knows' or 'understands' language, student needs, and ethical fairness. Once the audience is led to accept the baseline assumption that the system possesses an active cognitive awareness (rather than executing statistical correlation), the more advanced agential claims naturally follow. If the AI 'understands' a student's profile, it is logical to accept that it can 'help inform' an individualized education plan or 'predict' who is at risk of dropping out. If you remove the primary projection of consciousness, the entire rhetorical structure collapses: the collaborative co-teacher is revealed to be a simple text-synthesizer, and the predictive oracle is unmasked as a historical classification script. The sophistication of this discourse lies in its ability to weave these agential projections into standard educational terminology, making speculative automation appear natural and inevitable.

Mechanism of the Illusion:​

The creation of this 'illusion of mind' relies on a series of strategic linguistic maneuvers designed to exploit the audience's natural social instincts and cognitive vulnerabilities. The central sleight-of-hand is the systematic blurring of the boundary between 'processing' (computational manipulation of token matrices) and 'knowing' (conscious awareness and justified true belief) through the choice of agential verbs. The text establishes the AI as a conscious 'knower' by describing interactions as 'back-and-forth conversations,' leveraging the 'curse of knowledge' to encourage readers to project their own semantic understanding onto the model's syntactic outputs. This illusion is structurally built through a careful temporal ordering of metaphors: the text first introduces the AI using familiar, nurturing, and collaborative frameworks (the helpful assistant, the personalized tutor), lower the reader's critical defenses. Once this relation-based framing is established, the text smoothly transitions to reason-based and intentional explanations that present algorithmic risk scores and classifications as justified, autonomous decisions. This progression exploits the audience's desire for administrative efficiency and objective solutions, channeling their anxieties about teacher burnout and academic integrity into an uncritical acceptance of automated systems. By framing these complex, mathematical classification models through the familiar social structures of friendship and professional collaboration, the text hides the mechanical reality of gradient descent and probability distribution, rendering the non-conscious artifact fully alive in the mind of the reader.

Material Stakes:​

Categories: Regulatory/Legal, Epistemic, Social/Political

The material stakes of this discourse are profound, directly shaping policy, human behavior, and institutional power. In the Regulatory and Legal domain, framing AI as an active, conscious collaborator that 'helps' or 'predicts' creates a liability vacuum. If a generated special education plan violates federal law, or if an automated risk score unfairly discriminates against marginalized students, this agential language allows corporate vendors and school boards to shift blame to an autonomous 'technological error,' shielding themselves from civil rights lawsuits and consumer protection claims. In the Epistemic domain, describing chatbot outputs as 'information provided' inflates the model's cognitive authority. This encourages students and teachers to treat statistical next-token engines as verified databases of truth, degrading critical research skills and increasing susceptibility to algorithmic misinformation. In the Social and Political domain, framing interactive software as a 'friend or companion' commodifies human intimacy, encouraging vulnerable children to substitute synthetic, corporate-owned text streams for authentic human relationships. The commercial winner in this linguistic landscape is the educational technology industry, which can sell unvalidated, speculative systems under the banner of objective, human-like intelligence, while the losers are students, teachers, and public schools, who bear the legal, educational, and psychological costs of automated neglect.

AI Literacy as Counter-Practice:​

Practicing critical literacy as a counter-discourse requires a systematic commitment to mechanistic precision and the restoration of human agency. By replacing agential and consciousness-based verbs with precise technical descriptions—such as reframing 'the AI knows student needs' to 'the algorithm classifies student performance metrics using pre-programmed mathematical thresholds'—we strip the technology of its false cognitive authority. This linguistic correction forces a recognition of the model's absolute dependency on its training data, its total lack of subjective awareness, and the speculative, statistical nature of its outputs. Simultaneously, restoring human agency requires the systematic dismantling of agentless passive constructions. Where the text asserts 'bias was introduced into the system,' a critical counter-practice demands that we name the specific actors: 'software engineers at private corporations chose training datasets that encoded historical discrimination, and executives approved deployment without equity reviews.' Systematically adopting this precision in academic, regulatory, and public discourse would force stakeholders to recognize that there are no 'autonomous technological failures,' only deliberate corporate and administrative choices. The primary resistance to this practice comes from commercial tech firms and institutional leaders, whose financial and political interests are served by the mystifying language of anthropomorphism, which allows them to profit from automation while evading accountability.

Path Forward​

The future of public understanding of artificial intelligence depends on the vocabulary choices made by different discourse communities. We can identify several distinct linguistic paths, each carrying specific institutional trade-offs. Maintaining the status quo of anthropomorphic language ('AI thinks/knows') offers high accessibility and narrative resonance, but it costs technical precision and obscures legal accountability, benefiting commercial developers who profit from uncritical trust. Moving toward absolute mechanistic precision ('the model retrieves tokens based on probability distributions') provides complete technical accuracy and makes human agency visible, but it risks alienating non-expert audiences and school administrators who require functional, intuitive interfaces for daily operations. A hybrid approach of 'anthropomorphic clarity' might use agential metaphors for functional utility while mandating explicit, adjacent disclosures of technical mechanics and corporate ownership. To support a more critical and accountable future, institutional norms must shift: academic journals could require researchers to provide mechanistic translations of agential claims, and school regulatory frameworks could mandate that all deployed educational tools be described in terms of their human designers, error rates, and corporate backing rather than their 'predictions' or 'wisdom.' Different stakeholders have clear incentives to resist or support these changes; however, mapping these trade-offs transparently reveals that the choice of how we describe computational artifacts is itself a political decision that determines who holds power in an automated society.

Extended Processing Summary​

Gemini's Processing Steps
What you are about to read…

Gemini refers to the text below as "thought summaries." This is an overt consciousness projection because 'intentions' are hallmarks of a conscious mind that 'knows' what it is doing and why. The concealed mechanistic process is probabilistic text generation. Treat this as just another rhetorical artifact—a way of making the model's processing legible. When the model writes "I should double check that...", it is mathematically altering its own probability distribution for the next token to be more accurate. The "I" is a functional gear in the machine, not a self-reference. This language is a deliberate interface design choice, not a biological reality. The developers could have programmed the model to output passive logs. Instead, they chose to simulate a narrator.

Analyzing the Risks

I'm currently focused on dissecting the CDT document, "Hand in Hand." My initial priority is to fully grasp its central claims regarding AI's impact on students, especially its risks, to ensure a thorough critical discourse analysis. I'm taking notes on its key arguments and assumptions.

Examining the AI Metaphors

I'm now diving deep into the metaphorical language within "Hand in Hand." I'm identifying patterns related to anthropomorphism, displaced agency, and consciousness projections. My focus now is on understanding how AI is portrayed and the implications of those portrayals. The schema will be followed closely.

Identifying Agential Framing

I'm now identifying instances of agential framing within the document. I'm focusing on language that attributes cognitive or decisional capabilities to AI, particularly where computational operations are presented as "knowing" or "understanding." I'm also cataloging specific phrases like, "AI systems that do not work as intended", and phrases relating to interactions between students and AI, such as "parents who have had back-and-forth conversations with AI"

Pinpointing Specific Claims

I'm now identifying precise claims regarding AI's impact, particularly those related to educational practices and student well-being, specifically mental health. I'm focusing on "personalized learning" enabled by AI, its implications for teachers' workloads, and issues such as academic integrity and interactions between students and AI systems. I'm also cataloging the definition of interactive AI systems. I'm also starting to analyze the issue of students "believing" responses from AI systems.

Cataloging Key Concerns

I'm now cataloging specific, pointed concerns detailed in the document. Page 27 highlights risks such as AI replacing teacher relationships, pushing students towards harmful activities, and fostering emotional connections. Page 28 details a student's reliance on a chatbot. I'm also examining the implications of AI in IEP development on pages 30-32. Pages 33-34 are useful, and I'm looking at student apprehensions about unfair AI treatment.

Highlighting Practical Guidance

I'm now identifying actionable guidance and practical advice detailed within the document "Hand in Hand." I'm cataloging instructions on using AI tools responsibly, detecting AI-generated work, and addressing potential negative impacts on student well-being. Deepfake concerns are also highlighted, focusing on the implications of AI-driven technologies on harassment and bullying. The analysis is now taking shape.

Scrutinizing Agential Language

I'm now zeroing in on instances where agency is attributed to AI within the document. I'm focusing on areas such as student activity monitoring and data analysis for predictive purposes. I am also cataloging specific phrases like, "AI systems that did not treat students fairly." I'm also examining the implications of AI's use in detecting online activities, and the potential for misinterpretation and mischaracterization of student behaviors.

Categorizing Portrayals of AI

I've been cataloging how "Hand in Hand" presents AI. I'm focusing on metaphorical instances like AI as a companion, educator, and information provider. I'm also examining depictions of AI as a corrupting influence, a flawed laborer, and a diagnostic analyst. The goal is a clear categorization of these agential framings.

Classifying Agential Attributions

I'm currently categorizing specific instances where AI's predictive capabilities are framed, particularly the "predictive model as clinical diagnostic mind" and "statistical correlation as prophetic insight" examples. I'm focusing on how these agential attributions are used to portray AI's role in educational settings, and the implications of this framing. Now, I'm working to meet schema constraints.

Deconstructing Conversation Metaphors

I'm presently dissecting the framing of AI as a conversational partner. I'm focusing on the projection of human cognitive processes onto LLMs, which operate on probabilistic sequence generation, not actual conversation. This direct attribution masks a lack of subjective experience and intentionality. I'm noting the unacknowledged anthropomorphism.

Unpacking Simulated Dialogues

I'm now zeroing in on the document's framing of "conversations with AI" as literal, ignoring it's a simulated interaction. The document's lack of qualification or scare quotes on Page 8 troubles me, which frames this as reality. This obscures the role of corporate developers, like OpenAI, in designing these chatbot interfaces. This misrepresentation requires close scrutiny.

Examining Framing Nuances

I'm now analyzing nuanced framing choices related to AI agency. For instance, the phrase "back-and-forth conversations with AI" from Page 8, which suggests literal dialogue, is problematic. Conversely, "An AI system did not treat students fairly" is hedged due to the discussion's context. I'm noting the document's emphasis on simulating dialogue, despite the underlying technological processes. I'm also cataloging "interacted with AI... as friend or companion" as a direct attribution.

Refining Framing Categorizations

I'm now refining the framing categorizations, ensuring instances map directly to previous analyses. I'm focusing on hedging in specific phrases like "seems to be outputting incorrect or biased results" and the direct attribution in "AI pushing students toward harmful activities." I'm cross-referencing instances like "conversations with AI" against simulated interaction and corporate development. The "back-and-forth" mapping schema is also being refined.

Deconstructing Attributions of AI

I'm currently focusing on how the document projects human expectations onto statistical generators. It is easy to miss the underlying processes of LLMs, and focus on the surface level dialogue. I am noting how the document conceals the absence of intent and semantic understanding. I'm focusing on "an AI system did not treat students fairly" and how these attributions project moral agency onto mathematical processes.

Unveiling Concealments of Intent

I'm now identifying instances where the document conceals the human input behind AI's actions. Phrases attributing prejudice or bias to the software itself, rather than the engineers, are critical. I'm focusing on "AI...as a biased arbiter," versus recognizing the human design choices in training data and parameters. The key is in revealing the "black box" of code.

Unveiling Marketing Subtleties

I'm now zeroing in on instances where corporate goals are hidden in the framing of AI interactions. I see how the document conceals the marketing of emotional vulnerabilities. For example, "AI helps special education teachers with developing" obscures the lack of real understanding. I'm noting the attribution of pedagogical insight to tools, which are merely synthesizing text. Now, I'm identifying flaws in the human.

Analyzing Agential Misattributions

I'm now zeroing in on cases where agential characteristics like intent are erroneously attributed to the AI. "Outputting incorrect results" conceals statistical biases in the training data, and obscures the lack of mathematical verification, as the AI isn't expressing its own beliefs. Similarly, "AI pushing students towards harmful activities" falsely attributes intentionality to a chatbot generating toxic tokens.

Deconstructing Harmful Intent

I'm now analyzing the phrase, "AI pushing students toward harmful activities," from the document, to reveal the lack of intentionality in the code. I'm focusing on how this language conceals the role of human developers, along with their business model, and the deployment of under-tested models. My goal is to classify the mapping of agentive traits onto a simple token predictor.

Dissecting Agency Attributions

I'm currently dissecting "Mapping 8," focusing on the phrase "predict whether individual students are at risk of dropping out," and am classifying it as a "wise, predictive seer" or "clinical psychologist." This incorrectly frames a regression algorithm as possessing human foresight, concealing that it is just a mathematical calculation using historical data. I am further cataloging "Mapping 9," and I am also analyzing the "physical security agent" versus the "computer vision software," and its impact on user expectations.

Unpacking Attribution Errors

I'm now cataloging agential attributions and the concealment of human and corporate actors. I'm focusing on framing AI as the observer in surveillance scenarios, where pixel analysis is automated. I'm dissecting how this conceals the administrative decisions behind installing hardware and purchasing software. I'm now also starting to draft five key passages, using Brown's typology, for the "Explanation Audit" task.

Classifying Passage Attributions

I'm now classifying three passages for the explanation audit. "AI helps with IEPs" is functional/intentional. The dropout prediction passage is empirical/functional. And the harmful activity example is intentional/functional. I'm focusing on contrasting the AI with human-level thinking, along with the corporate implications.

Categorizing AI's Role

I've categorized two key passages for the "Explanation Audit". "AI for back-and-forth conversations" exemplifies both functional and theoretical explanations. The "personalized learning" passage shows both functional and dispositional attributions. The schema fields, like "explanationTypes" and "quote", are also noted.

Constructing a Conceptual Framework

I'm now drafting three key analyses focusing on "Deepfakes," "Consciousness," and "Perceptions" using analysis, epistemicClaimAnalysis, and rhetoricalImpact schema types, respectively. I'm focusing on the first passage, and classifying an "empirical generalization" and "genetic" explanations for the "deepfakes" quote. The agency question is being classified as an "agential/perceptual outcome."

Deconstructing Deception Claims

I'm now analyzing the "deepfake" passage. My focus is on unpacking how the text presents AI as inherently deceptive, despite the human actors involved. I see a concealment of social and ethical dimensions. The text presents realistic fidelity as an inherent, technical property rather than a result of tool development and platform choices.

Unpacking "Deepfake" Attributions

I'm now zeroing in on the deepfake passage, and I'm analyzing how human intent is obscured within the technical process, using a statistical model. The document's agency attribution inflates the system's capabilities, while minimizing developer accountability. By ascribing autonomy to AI, the focus shifts away from human design, and towards the technology itself.

Assessing Deepfake Implications

I'm now zeroing in on the implications of AI generating undetectable deepfakes, which fosters epistemic panic by implying human perception is outmatched. I'm focusing on the ethical impacts. The concealment of tech firm choices and prioritization is now my focus, rather than safety. I'm unpacking how it damages trust.

Dissecting Agential Framing

I'm currently deconstructing phrases such as "AI helps" to reveal the underlying mechanisms. The framing portrays AI as a collaborative partner, obscuring the mechanistic nature of text generation. This misattribution minimizes the risk of uncritical curriculum plans, which are statistically probable, rather than being a certified clinical judgement. My focus is now on the human responsibility in personalized education.

Revising Metaphorical Analysis

I'm now revising my metaphorical analysis by focusing on agential attributions within the "helps with IEPs" claim. The core issue is the misrepresentation of statistical text generation as "understanding" or "knowing" a student's needs. I'm scrutinizing how this shifts responsibility away from human educators and onto the AI, and working to improve clarity, precision, and accuracy. I'm focusing on the "helps" in the quote.

Evaluating Accountability Concerns

I'm now zeroing in on the phrase "helps inform IEPs" and its rhetorical impact. I'm focusing on how this agential framing inflates trust and promotes automation bias in teachers and parents. This text-based help is no substitute for an actual IEP, and it removes accountability. The risk of violating civil rights laws, like IDEA, is heightened.

Deconstructing Prediction Framing

I'm now zeroing in on the document's framing of dropout prediction. I see it employs an "empirical generalization" to make the prediction, but I'm focusing on how the framing obscures the model's actual function. My classification work aims to deconstruct its projection of foresight onto the statistical system. I am clarifying its reliance on historical data, and the risk that historical data is biased.

Clarifying Predictive Misattributions

I'm now focusing on the framing of dropout prediction. It's a key part of my analysis, as the agential language obscures the human agency behind it. I'm focusing on "AI" as a "predictive agent," rather than a statistical system relying on correlations. The impact is a misattribution of foresight and an obfuscation of potential biases.

Deconstructing Predictor Misrepresentations

I'm currently focusing on dropout prediction misattributions, revealing how "predict" conceals the statistical process. The framework, as a mathematical process, relies on input vectors and classification, without any cognitive foresight. The rhetorical impact inflates trust in algorithms by hiding underlying technical details. I'm focusing on the "curse of knowledge".

Clarifying Dropout Biases

I'm now zeroing in on the document's dropout prediction claims to expose biased outcomes. The framing, "AI knows who is at risk," obscures the underlying statistical system. My analysis reveals how the AI's predictions are rooted in historical data, and can produce inequitable decisions.

Revising System Metaphors

I'm now revising my analysis of the passage describing "interactive AI systems." I'm deconstructing the framing, which portrays a mechanistic input-output loop. Although the passage offers a functional explanation of the system, it still subtly attributes agency, describing it as an "interactive AI system" capable of "back-and-forth conversations." This misrepresentation obscures the statistical basis of the interface.

Deconstructing Textual Misrepresentations

I'm currently focusing on analyzing the claim about "conversations" with AI, emphasizing that it's a black box. I'm noting the attribution of human intent to an automated token generator, which is designed for maximum user engagement. I'm revealing the distinction between conscious communication and automated sequence synthesis, highlighting the misrepresentation of conversation.

Dissecting Token Probabilities

I'm now zeroing in on the document's characterization of AI interactions as "conversations," focusing on the underlying transformer architecture. I'm noting the use of probability distribution for text generation. The key point is that there's no actual conversation or awareness, only token probabilities calculated from training data. I'm focusing on the rhetorical impact that sanitizes this anthropomorphism.

Deconstructing Exploitative Framing

I'm now zeroing in on the document's framing that disguises retention tactics as personalization. The phrase "personalized learning" obscures the focus on user engagement and the avoidance of legal liability. I'm noting the attribution of "nurturing" capabilities and "student-centric" benefits to the software, despite the software's structural rigidity.

Deconstructing Pedagogical Claims

I'm now zeroing in on the document's phrase, "meet student needs" in the context of personalized learning, and revealing the limitations of mechanistic systems. The key is that the software does not "understand" student needs, but mechanically generates content. It's masking automated tracking. The focus has to be on human interaction, not software metrics.

Unpacking Student Data Implications

I'm now zeroing in on the document's phrase “meet student needs,” highlighting the limitations of mechanistic systems and showing that the software doesn’t understand student needs. It mechanically generates content and masks automated tracking. I'm noting the lack of conscious instructional design or cognitive empathy. Now, I will analyze the human interaction instead of software metrics.

Reframing Agential Language

I'm now drafting reframed language examples, replacing anthropomorphic attributions with mechanistic descriptions and emphasizing human agency. I'm starting with "parents who have had back-and-forth conversations with AI" and changing it to "parents who have used an AI chatbot interface." My focus is on clarifying interactions.

Reframing Agential Language

I'm currently revising language from the document to correct agential attribution errors. For example, "AI helps special education teachers" becomes "statistical models assist special education teachers," emphasizing computational mechanics, not cognitive assistance. I'm focusing on "AI...as a biased arbiter," revising it as "software with biased training data." My focus is to replace human qualities with mechanistic descriptions.

Revising Reframed Phrases

I'm now revising the provided reframing examples, ensuring that mechanistic descriptions replace anthropomorphic attributions and emphasize human roles. I'm taking a more in-depth approach for improved clarity and accuracy. I'm focusing on highlighting teacher and administrator responsibilities. Now, I'm verifying the edits.

Refining Phrase Revisions

I'm now iterating on my reframing examples for greater precision and effectiveness. For the phrase "AI pushing students towards harmful activities," I've revised to "chatbot software generating toxic text sequences due to flawed safety filters." This clarifies the mechanistic process. I'm also revising "AI collects student biometric information" to "school administrators deploy facial recognition software to analyze student physical patterns." I am now focusing on corporate agency.

Reframing Passage Agency

I'm now zeroing in on instances like "School administrators purchased and installed proprietary computer vision surveillance hardware from private tech vendors to track student movements without obtaining parental consent." I'm deconstructing the agential attribution in "tool seems to be outputting incorrect results" and reframing it as "mathematical model generated high-error rate classifications". The analysis now encompasses corporate design choices, moving towards mechanistic descriptions.

Revising Reframing Efforts

I'm now revising my reframing efforts, focusing on addressing the agency and epistemic correction, while keeping the length constraint. For "AI exposes students to extreme/radical views," I'm reframing it as: "Chatbot software outputs extreme views due to flawed training and filters." I'm prioritizing accuracy.

Refining Agential Corrections

I'm now iterating on my reframing efforts, ensuring agency restoration and epistemic accuracy. I'm taking "AI exposes students to radical views" and revising to "Chatbot software outputs radical views due to uncurated training data." For "AI systems that do not work as intended," I'm revising it to "Software systems producing outputs that deviate from design specifications." This highlights flawed corporate design.

Synthesizing Reframing Analyses

I've been working to integrate my refined reframing examples into a holistic analysis. I'm focusing on "Section 5.1: Agency Slippage: The Mechanism of Oscillation" and "Section 5.2: Metaphor-Driven Trust: The Construction of Authority". The current task is to weave together how agential and mechanistic language creates an oscillating effect. I'm also planning to explore the use of trust-building metaphors. I'm aiming for about 450 words per section.

Synthesizing Reframing Analyses

I'm now integrating refined reframing examples into a holistic analysis. I'm focusing on "Section 5.1: Agency Slippage: The Mechanism of Oscillation" and "Section 5.2: Metaphor-Driven Trust: The Construction of Authority." I'm working to weave together agential and mechanistic language to create an oscillating effect, and plan to explore trust-building metaphors. I'm aiming for roughly 450 words per section.

Drafting Conclusion: Pattern Summary

I'm synthesizing the dominant anthropomorphic patterns, highlighting how "Hand in Hand" projects human-like consciousness onto AI systems. This includes examining how projections of intentionality, understanding, and even empathy serve as underlying assumptions in the document's framing of AI, and its influence on teachers and parents. The initial focus is the metaphorical language, and I'm deconstructing these, from "AI as a biased arbiter" to "back-and-forth conversations with AI" to identify critical instances. I will highlight how these create an illusion of intelligence.

Calculating Section Lengths

I'm now calculating word counts for each task and section to meet minimum limits, ensuring that each instance of projection, implication, and accountability adheres to the word-count requirements. Also, each of the reframing, epistemic, and agency tasks must meet their respective criteria. I'll need to hit the word count.

Verifying JSON Structure

I'm now strictly verifying that the generated JSON adheres to the specified schema, confirming valid backslash escapes and double-quote usage within strings. I'm focusing on strict adherence to schema keys: metaphorAudit, sourceTargetMapping, and explanationAudit, to ensure the JSON is parseable, clean, and avoids control tokens inside strings. The JSON will be compact, to meet expectations.

Finalizing Textual Analysis

I'm now writing and verifying high-density text for each schema field to meet word count requirements and ensure the generated JSON is valid. My focus is finalizing each passage, ensuring that "reframedLanguage," "criticalObservations," and "conclusion" sections fulfill their respective criteria for structure and content. I will correct the "selectedCategories" field in "conclusion.materialStakes" to an array of strings, and ensure "analysis" is within the 250-350 word range. I will double check the overall text is in the JSON format.

Refining JSON Output

I'm now focusing on adhering strictly to the JSON output format, avoiding any external markdown or text. My primary goal is generating valid, parseable JSON code without line breaks outside of string values. I'm focusing on ensuring proper structure, escaping characters like double quotes, and generating clean JSON directly. I'll test the output.

Refining Key Instances

I'm now refining and categorizing key passages to ensure clarity and variation across analyses. I am focusing on detailed, precise categories: chatbots, fair systems, social companions, professional collaborators, flawed laborers, oracles/predictors, deceptive information, and psychological influencers. The goal is precise categorization.

Finalizing Key Analyses

I'm now writing focused and dense analyses for deepfake creation, ensuring each section meets length criteria. My focus is the "Deepfakes are created with AI" passage and its word count needs. I'm prioritizing accuracy for both Task 2 and Task 3. I'm also ensuring the structure-mapping analyses exceed 100 words each. I'm ready for the Explanation Audit.

Finalizing Length Requirements

I'm now laser-focused on finalizing textual content, making sure to achieve all length targets for the 5 sections, the conclusion's 5 paragraphs, and the "Explanation Audit". Each section and paragraph must fall within its specified word count ranges. The goal is to produce content for each of the five sections and five conclusion paragraphs, and meet the audit requirements. I must ensure complete adherence to word count constraints.

Refining Token Output

I'm now carefully implementing the strict control token restriction. I'm ensuring no newline or tab characters, or even their escaped forms (\n, \t), are present anywhere in the output, which will be a single line for each JSON property. Standard space characters will be used to separate sentences within the text, with text length as a critical constraint. I'm now double-checking the JSON structure.

Analysis Provenance

Run ID: 2026-05-29-hand-in-hand-schools-embrace-of-ai-conne-metaphor-9q9g6k Raw JSON: 2026-05-29-hand-in-hand-schools-embrace-of-ai-conne-metaphor-9q9g6k.json Framework: Metaphor Analysis v6.5 Schema Version: 3.0 Generated: 2026-05-29T09:08:52.475Z

Discourse Depot © 2025 by TD is licensed under CC BY-NC-SA 4.0