Skip to main content

Artificial Intelligence Cognition and Societal Problem-Solving: A Theoretical and Computational Examination of Machine Thinking, Operational Logic, and Applied Intelligence in Contemporary Society

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. Machine as Independent Cognitive Entity​

Quote: "This study examines how AI "thinks," performs operations, and exhibits cognitive-like abilities in solving real-world problems"

  • Frame: System as thinking organism
  • Projection: The metaphor maps the distinctly human capacity for conscious thought, reflection, and subjective deliberation onto automated computational processes. By suggesting the system 'thinks' and exhibits 'cognitive-like abilities,' the text projects the possession of an internal mental life, contextual awareness, and the capacity for justified true belief onto algorithmic operations. Human thinking involves continuous subjective experience, an understanding of meaning, and the ability to evaluate the truth value of propositions. In contrast, AI systems exclusively process numerical weights, execute token prediction, and perform mathematical correlations within high-dimensional vector spaces. Projecting cognitive abilities onto these mechanisms fundamentally conflates mechanistic pattern-matching with conscious understanding, creating an illusion of mind where there is only mathematical optimization and statistical probability distribution.
  • Acknowledgment: Hedged/Qualified (The author places "thinks" in quotation marks and uses the qualifier "cognitive-like abilities." I considered Explicitly Acknowledged, but the immediate context does not offer explicit meta-commentary on the metaphor itself; it merely hedges the language. The quotes signal a non-literal use without unpacking the underlying mechanistic reality.)
  • Implications: Framing an artificial intelligence system as an entity capable of thinking and cognitive problem-solving significantly inflates public perception of the technology's sophistication. When stakeholders believe a system is 'thinking,' they are inclined to extend relation-based trust to its outputs, assuming the machine has evaluated the context and truthfulness of its generated answers. This unwarranted trust creates substantial risks in domains like healthcare or criminal justice, where users may defer to an algorithm's statistical correlation as if it were a considered, logical judgment, masking the brittleness of the underlying pattern recognition.

Accountability Analysis:

  • Actor Visibility: Hidden (agency obscured)
  • Analysis: The text attributes the action entirely to 'AI' as the subject that performs operations and solves problems. I considered Ambiguous, but the grammatical structure clearly positions the AI as an independent actor. This agentless construction obscures the human researchers, corporate developers, and deployment teams who actually define the problem parameters, select the training data, and program the operational logic. By portraying the machine as the solver of real-world problems, the text shields the designers from accountability for the specific, value-laden choices embedded within those computational processes.
Show more...

2. Algorithmic Operation as Human Reasoning​

Quote: "Through algorithms and data-driven models, AI systems perform operations that mimic reasoning, learning, and decision-making"

  • Frame: Processing as cognitive reasoning
  • Projection: This framing maps the human intellectual practices of reasoning, learning, and decision-making onto the mechanistic execution of algorithms. Human reasoning requires conscious logical deduction, semantic comprehension, and the weighing of abstract concepts. Human learning involves the contextual assimilation of knowledge and meaning. Conversely, the AI system being described merely adjusts mathematical weights across neural network layers to minimize error rates (backpropagation) and calculates statistical probabilities based on training data. By projecting 'reasoning' and 'learning' onto the system, the text implies that the machine understands the logical relationships between variables and consciously acquires knowledge, rather than blindly executing mathematical optimization functions without any awareness of the data's real-world significance.
  • Acknowledgment: Hedged/Qualified (The text uses the verb 'mimic' to qualify the actions, indicating that the system's operations resemble but do not genuinely equate to human reasoning. I considered Direct (Unacknowledged) because 'learning' is often used literally in computer science, but 'mimic' modifies the entire list, functioning as a clear hedge recognizing the simulation.)
  • Implications: Even when hedged with 'mimic,' applying terms like reasoning and decision-making to algorithms drastically alters policy and governance frameworks. It encourages policymakers to treat algorithms as autonomous decision-makers rather than static mathematical tools applied by human institutions. This can lead to the implementation of automated systems in sensitive social areas under the false assumption that the system can 'reason' through edge cases or ethical dilemmas, ultimately resulting in brittle and potentially harmful applications of statistical models that completely lack the capacity to reason about human context.

Accountability Analysis:

  • Actor Visibility: Hidden (agency obscured)
  • Analysis: The AI systems are positioned as the grammatical subjects performing the mimicking and decision-making. I considered Partial visibility because the text mentions 'algorithms and data-driven models' as the means, but it fails to identify the human actors who create and deploy these models. The engineers who code the algorithms and the executives who mandate their use are entirely absent. This displacement serves institutional interests by transferring the agency of complex organizational decisions to a supposedly neutral, reasoning technological artifact.

3. Data Processing as Social Comprehension​

Quote: "there is insufficient attention to how AI systems interpret and respond to complex social dynamics"

  • Frame: Algorithm as social interpreter
  • Projection: The text maps the deep human capacity for social interpretation, empathy, and cultural understanding onto a mathematical system. When humans 'interpret' social dynamics, they utilize a theory of mind, cultural context, emotional intelligence, and lived experience to grasp nuanced interactions. Projecting this onto AI attributes conscious semantic understanding to a system that only processes numerical embeddings. The machine does not 'interpret' society; it classifies tokens, extracts features, and generates statistical predictions based on patterns in its training corpus. Using verbs like 'interpret' and 'respond' implies the AI possesses a subjective viewpoint and the capacity to comprehend the meaning of social phenomena, rather than merely calculating mathematical distances between data points.
  • Acknowledgment: Direct (Unacknowledged) (The text presents the AI's ability to 'interpret and respond' as a literal, factual capability requiring more attention. I considered Hedged/Qualified, as the broader paper disclaims AI consciousness, but this specific assertion is presented entirely without qualification or quotation marks, framing the interpretive capacity as an established fact.)
  • Implications: Attributing the capacity for social interpretation to AI fundamentally misrepresents the nature of computational bias. It leads audiences to believe the system actively understands and navigates social complexities, masking the reality that the system is merely reproducing historical correlations found in its training data. This overestimation of capability can lead to catastrophic deployments in social services and predictive policing, where mathematical feedback loops are falsely validated as deep, responsive social insights, thereby laundering historical discrimination through the illusion of objective machine comprehension.

Accountability Analysis:

  • Actor Visibility: Hidden (agency obscured)
  • Analysis: The text isolates 'AI systems' as the sole active entities interpreting social dynamics. I considered Named because researchers are implicitly the ones giving 'insufficient attention,' but regarding the actual action of the system, agency is completely displaced. It obscures the human developers who label the data, the sociologists who operationalize the variables, and the institutions that define what constitutes a 'response.' This framing absolves the human creators of the responsibility for how their statically trained models impact fluid social environments.

4. Pattern Adjustment as Autonomous Adaptation​

Quote: "reinforcement learning enables AI systems to make sequential decisions by maximising cumulative rewards"

  • Frame: Optimization as autonomous decision-making
  • Projection: This metaphor projects the human experience of conscious choice, goal-oriented intent, and deliberate decision-making onto the mechanistic process of algorithmic optimization. Human decision-making involves evaluating options, understanding consequences, and exercising agency based on internal desires or ethical frameworks. In reinforcement learning, the system mathematically updates its policy to increase a numerical scalar (the reward signal) based on an externally programmed objective function. The text projects conscious agency onto the algorithm, suggesting it 'makes decisions' and actively 'seeks' rewards. This conceals the fact that the machine has no awareness of the decisions, no comprehension of the reward, and operates strictly according to deterministic or stochastic mathematical updating rules governed by human-defined parameters.
  • Acknowledgment: Direct (Unacknowledged) (The explanation is delivered as a straightforward technical description of how reinforcement learning functions, with no hedging around 'make sequential decisions.' I considered Hedged/Qualified because it mentions the mechanism ('maximising cumulative rewards'), but the agential verb 'make' is presented as literal fact.)
  • Implications: By framing mathematical optimization as autonomous decision-making, the text normalizes the delegation of critical operational choices to algorithms. It creates an aura of strategic brilliance around the system, implying it possesses strategic foresight. This obscures the fragility of reinforcement learning, particularly its tendency to exploit poorly specified reward functions in unintended ways (reward hacking). When audiences believe the system is making rational decisions, they are less likely to implement strict human oversight or critically audit the human-designed reward functions that actually drive the behavior.

Accountability Analysis:

  • Actor Visibility: Partial (some attribution)
  • Analysis: The text explicitly notes that 'reinforcement learning enables' the system, and earlier context mentions predefined objectives, implying human design. However, I considered Hidden because the human engineers who define the reward function and constrain the environment are not explicitly named in this construction. While partial mechanisms are visible, the active agency is still heavily displaced onto the 'AI systems' making decisions, deflecting attention from the human actors who determine exactly what constitutes a 'reward' and whose interests that reward serves.

5. Opacity as Inherent Mental Mystery​

Quote: "The opacity of machine learning models limits transparency and accountability in decision-making processes. This is particularly problematic in high-stakes domains"

  • Frame: System complexity as autonomous mystery
  • Projection: This framing projects the inherent mystery of the human mind—the inability to perfectly access another's subjective thoughts—onto the technical phenomenon of computational opacity. By characterizing the model as an opaque 'black box' making decisions, the text implies that the system possesses internal, inaccessible cognitive states akin to a human mind that cannot be fully interrogated. This maps the natural limits of interpersonal understanding onto a designed artifact. It obscures the fact that the 'opacity' is a result of mathematical dimensionality, proprietary corporate secrecy, and deliberate engineering choices to prioritize predictive accuracy over interpretability, rather than an inherent, organic mystery of a conscious entity.
  • Acknowledgment: Direct (Unacknowledged) (The opacity is stated as a literal, inherent property of the machine learning models. I considered Explicitly Acknowledged because the text uses the term 'black box' elsewhere, but in this specific quote, the opacity is presented straightforwardly as an independent characteristic of the model that actively 'limits transparency'.)
  • Implications: Framing opacity as a natural, unavoidable characteristic of the system rather than a human design choice creates a powerful 'accountability sink.' If the system is inherently mysterious, like the human mind, then developers and corporations cannot be expected to fully explain its outputs. This framing serves as a pre-emptive defense against regulation, suggesting that transparency is biologically or technically impossible rather than commercially inconvenient. It fosters a regulatory environment where 'we don't know how it works' is accepted as a valid excuse for deploying harmful systems.

Accountability Analysis:

  • Actor Visibility: Hidden (agency obscured)
  • Analysis: The phrase 'The opacity of machine learning models limits transparency' transforms a structural design choice into an independent actor. I considered Partial visibility because the text discusses accountability, but no specific humans or corporations are named. It hides the specific technology companies, corporate executives, and software engineers who deliberately choose to deploy high-dimensional neural networks instead of interpretable models (like decision trees) because they prioritize performance and proprietary lock-in over public accountability and transparency.

6. Statistical Output as Independent Knowledge Generation​

Quote: "AI contributes to crime prevention through predictive policing algorithms. These applications demonstrate AI's capacity to process complex datasets and generate actionable insights"

  • Frame: Algorithm as knowledge generator
  • Projection: This metaphor maps the human intellectual process of deriving insights, understanding contexts, and generating novel knowledge onto the mechanistic process of statistical forecasting. A human generates an 'insight' through a sudden conscious realization or deep contextual comprehension of an underlying reality. The AI system, however, merely performs mathematical regression, finding correlations within historical data arrays to minimize prediction error. By projecting the capacity to generate 'actionable insights' onto the AI, the text attributes epistemic authority and conscious realization to a mathematical model. This obscures the fact that the machine has no comprehension of crime, society, or the human consequences of its statistical outputs; it strictly processes numerical correlates.
  • Acknowledgment: Direct (Unacknowledged) (The claim that AI generates actionable insights is presented as an objective, literal demonstration of its capabilities. I considered Hedged/Qualified since it limits the capacity to processing datasets, but the term 'insight' remains completely unhedged, granting the system literal epistemic agency.)
  • Implications: Using the term 'insights' to describe statistical correlations grants immense, unwarranted epistemic authority to predictive policing systems. It encourages law enforcement and policymakers to view algorithmic outputs as profound, objective truths rather than reproductions of historical arrest data, which often reflect systemic biases. This consciousness projection shifts the burden of proof; if the machine generates an 'insight,' humans must prove it wrong, rather than demanding the creators prove its validity. It provides a veneer of objective intelligence to discriminatory practices.

Accountability Analysis:

  • Actor Visibility: Hidden (agency obscured)
  • Analysis: The grammatical structure positions 'AI' as the entity that 'contributes' and 'applications' as demonstrating capacity. I considered Partial, as 'policing algorithms' implies a domain of use, but human agency is completely erased. The text obscures the police departments that purchase the software, the software companies (like PredPol or Palantir) that build it, and the human data entry clerks whose historical biases populate the datasets. By framing AI as the generator of insights, responsibility for the subsequent police actions is deflected onto the technology.

7. Mathematical Bias as Behavioral Misconduct​

Quote: "This limitation is evident in cases where AI produces biased or inappropriate outputs"

  • Frame: Algorithm as biased actor
  • Projection: This framing maps human social failings—prejudice, bias, and inappropriate behavior—onto mathematical functions. When a human is biased, it involves conscious or unconscious prejudice, subjective judgments, and a failure of moral reasoning. Projecting this onto an AI system suggests the system possesses its own agency and actively discriminates. In reality, the AI system strictly processes matrix multiplications based on weights derived from training data. It does not 'produce bias' through any internal prejudice; it mathematically replicates and amplifies the statistical distribution of the human-generated data it was fed. This projection grants the machine a false moral agency, making it seem like a misbehaving actor rather than a mirror reflecting societal data.
  • Acknowledgment: Direct (Unacknowledged) (The statement is presented as a literal, factual observation of the system's behavior. I considered Hedged/Qualified because it is framed as a 'limitation,' but there is no hedging around the claim that the AI is the active producer of the bias. It is stated as a direct technical reality.)
  • Implications: By characterizing the AI as the producer of bias, the discourse subtly shifts the focus of ethical intervention from human societal structures to technological patching. If the AI is biased, the solution appears to be 'fixing the algorithm' rather than addressing the structural inequalities encoded in the historical data or questioning the decision to use automated systems in that domain. This anthropomorphic framing protects the institutions providing the biased data and the companies building the models, as it conceptualizes the failure as a 'glitch' in the machine rather than a faithful reproduction of institutional prejudice.

Accountability Analysis:

  • Actor Visibility: Hidden (agency obscured)
  • Analysis: The AI is positioned as the active subject ('AI produces'). I considered Partial because the text elsewhere discusses data dependency, but in this specific formulation describing the problem, human actors are entirely erased. It hides the data brokers who compile the datasets, the human annotators who label them, and the engineering teams who fail to audit the data for representational harm. By grammatically attributing the production of bias to the AI, the text provides a linguistic shield for the human actors whose decisions actually determined the model's parameters.

8. System Dependency as Social Collaboration​

Quote: "This perspective aligns with the concept of augmented intelligence, which emphasises collaboration between humans and machines"

  • Frame: Algorithm as collaborative partner
  • Projection: This metaphor projects the dynamics of human social relationships, teamwork, and mutual understanding onto the interaction between a human user and a computational tool. Collaboration inherently requires two conscious agents possessing shared goals, mutual awareness, and the capacity to negotiate and adapt to one another's intentions. By framing the use of AI as 'collaboration,' the text attributes social agency, communicative intent, and a shared theory of mind to the algorithm. Mechanistically, the AI does not collaborate; it passively receives input prompts, processes tokens through attention layers, and outputs statistically probable responses. It has no awareness of the human, no shared goals, and no capacity for true collaborative negotiation.
  • Acknowledgment: Direct (Unacknowledged) (The concept of collaboration is presented as a straightforward description of the paradigm of 'augmented intelligence.' I considered Explicitly Acknowledged because it references a specific theoretical concept, but the relational quality of 'collaboration' between human and machine is stated as a literal, unhedged functional reality.)
  • Implications: Framing AI as a collaborator rather than a tool significantly alters human-computer interaction dynamics, fostering inappropriate emotional and professional reliance. When professionals view a system as a collaborator, they are more likely to succumb to automation bias, trusting the system's outputs as the contributions of a competent peer rather than critically evaluating them as statistical generations. This false sense of partnership obscures the reality that the 'collaborator' cannot take moral or legal responsibility for its errors, leaving the human user entirely liable for the unverified outputs of a mathematical model.

Accountability Analysis:

  • Actor Visibility: Hidden (agency obscured)
  • Analysis: While 'humans' are named as half of the partnership, the other half is abstractly termed 'machines,' establishing a false equivalence of agency. I considered Named because humans are explicitly part of the equation, but regarding the creation and constraints of the machine, the corporate actors are hidden. The framing obscures the reality that the user is not collaborating with a machine, but is actually interacting with the encoded decisions, guardrails, and business logic of the specific technology corporation that deployed the model. The 'collaboration' masks corporate mediation.

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: Conscious human thinker with internal mental states, cognitive processing, and subjective problem-solving abilities. → Computational system executing algorithmic operations, mathematical optimization, and statistical pattern matching.​

Quote: "This study examines how AI "thinks," performs operations, and exhibits cognitive-like abilities in solving real-world problems"

  • Source Domain: Conscious human thinker with internal mental states, cognitive processing, and subjective problem-solving abilities.
  • Target Domain: Computational system executing algorithmic operations, mathematical optimization, and statistical pattern matching.
  • Mapping: The structural mapping transfers the architecture of a conscious mind onto the architecture of a computer program. In the source domain, a thinker possesses intentionality, awareness of context, and an understanding of the semantic meaning of the problem being solved. This relational structure is mapped onto the target domain such that executing code is equated with 'thinking,' and producing a mathematically optimal output is equated with 'solving' a real-world problem. This mapping invites the assumption that the system possesses a subjective awareness of the data it processes and an intentional drive to achieve a resolution, transferring the epistemic weight of human consciousness onto mindless statistical correlation.
  • What Is Concealed: This mapping aggressively conceals the complete absence of semantic understanding within the AI system. It obscures the mechanistic reality that the system manipulates ungrounded symbols (tokens, vectors, pixels) based purely on syntactic rules and statistical proximity, without any connection to real-world meaning. It also hides the heavy reliance on human labor: the engineers who translate the 'real-world problem' into a mathematical optimization objective, and the human workers who manually label the data. It replaces a transparent view of algorithmic mechanics with an opaque illusion of mental activity.
Show more...

Mapping 2: Human learner and decision-maker capable of logical deduction, knowledge acquisition, and deliberate choice. → Machine learning processes, specifically backpropagation for weight adjustment and probabilistic classification.​

Quote: "Through algorithms and data-driven models, AI systems perform operations that mimic reasoning, learning, and decision-making"

  • Source Domain: Human learner and decision-maker capable of logical deduction, knowledge acquisition, and deliberate choice.
  • Target Domain: Machine learning processes, specifically backpropagation for weight adjustment and probabilistic classification.
  • Mapping: This mapping aligns human intellectual development with mathematical model fitting. In the source domain, reasoning involves connecting premises to conclusions through logic; learning involves integrating new concepts into a worldview; and decision-making involves weighing options against values. In the target domain, the AI updates numerical parameters to reduce error rates based on a loss function (learning), calculates statistical likelihoods (reasoning), and selects the output with the highest probability score (decision-making). The mapping invites the assumption that the model's internal operations follow logical, understandable paths similar to human thought processes, transferring the justification of human rationale onto probabilistic mechanics.
  • What Is Concealed: The mapping conceals the purely mathematical, non-conceptual nature of the target domain. It hides the fact that gradient descent and backpropagation do not involve understanding or logic, but rather blind mathematical optimization over a multi-dimensional error surface. It obscures the system's brittleness: unlike a human who reasons, an AI can fail catastrophically if input data slightly deviates from the training distribution (adversarial examples). By mapping cognitive verbs onto these processes, it conceals the profound differences between statistical correlation and causal human understanding.

Mapping 3: Human social actor possessing empathy, cultural awareness, and an active theory of mind. → Algorithmic classification and prediction models applied to sociological or demographic datasets.​

Quote: "there is insufficient attention to how AI systems interpret and respond to complex social dynamics"

  • Source Domain: Human social actor possessing empathy, cultural awareness, and an active theory of mind.
  • Target Domain: Algorithmic classification and prediction models applied to sociological or demographic datasets.
  • Mapping: This mapping projects the relational structure of social interaction onto data processing. In the source domain, a social actor perceives cues, understands cultural contexts, interprets implicit meanings, and formulates a measured, socially appropriate response. When mapped to the target domain, mathematical feature extraction becomes 'interpretation,' and statistical output generation becomes a 'response.' This invites the dangerous assumption that the algorithmic system possesses a conscious, nuanced understanding of societal complexities and can dynamically adapt its behavior based on a genuine comprehension of the human condition.
  • What Is Concealed: This mapping entirely conceals the static, backward-looking nature of the AI system. The target domain does not interact with fluid social reality; it processes frozen, historical data representations chosen by developers. The mapping hides the reality that what is called 'interpretation' is actually mathematical categorization based on proxy variables (e.g., using zip codes as a proxy for income). It obscures the profound transparency obstacle: the model's inability to explain why it made a correlation in terms that make social sense, hiding the corporate design choices behind a veil of perceived artificial wisdom.

Mapping 4: A rational, goal-oriented agent making strategic choices to maximize personal benefit or utility. → A Markov decision process algorithm updating its policy function via stochastic gradient descent based on a programmed scalar signal.​

Quote: "reinforcement learning enables AI systems to make sequential decisions by maximising cumulative rewards"

  • Source Domain: A rational, goal-oriented agent making strategic choices to maximize personal benefit or utility.
  • Target Domain: A Markov decision process algorithm updating its policy function via stochastic gradient descent based on a programmed scalar signal.
  • Mapping: This mapping aligns conscious strategic planning with programmatic policy updating. In the source domain, a rational agent assesses a situation, looks to the future, makes choices, and seeks a rewarding outcome based on desires. In the target domain, the algorithm explores a constrained mathematical environment, calculates expected values using the Bellman equation, and adjusts probabilities to maximize an externally defined numerical scalar. The mapping invites the assumption that the algorithm possesses foresight, desires, and autonomous agency, making it seem like an independent actor pursuing its own goals.
  • What Is Concealed: The mapping conceals the rigid, pre-programmed determinism of the reward structure. It hides the fact that the 'reward' is not a subjective experience of pleasure or success, but a literal integer value programmed by a human engineer. It obscures the phenomenon of reward hacking, where systems exploit loopholes in the mathematical environment without any 'understanding' that they are violating the spirit of the task. Crucially, it conceals the human agency behind the objective function, making the system's behavior seem like a natural expression of intelligence rather than the execution of human-coded parameters.

Mapping 5: An inherently mysterious, inaccessible natural phenomenon or subjective mind (the 'black box' of human consciousness). → High-dimensional neural networks with millions or billions of parameters, often protected by corporate trade secrets.​

Quote: "The opacity of machine learning models limits transparency and accountability in decision-making processes. This is particularly problematic in high-stakes domains"

  • Source Domain: An inherently mysterious, inaccessible natural phenomenon or subjective mind (the 'black box' of human consciousness).
  • Target Domain: High-dimensional neural networks with millions or billions of parameters, often protected by corporate trade secrets.
  • Mapping: This mapping projects the ontological mystery of human consciousness onto a designed computational artifact. In the source domain, one cannot directly observe the internal workings of another's mind; it is naturally opaque. When mapped to the target domain, the complex, highly non-linear matrix multiplications of a neural network are treated as similarly inscrutable. This mapping invites the assumption that AI opacity is an unavoidable law of nature or an inherent characteristic of intelligence, rather than a specific engineering consequence of choosing complex architectures over interpretable ones.
  • What Is Concealed: This metaphor powerfully conceals the economic and design choices that create the opacity. It hides the fact that transparency is often limited not just by mathematical complexity, but by intellectual property laws, corporate nondisclosure agreements, and a deliberate industry preference for highly parameterized predictive models over simpler, explainable algorithms. It conceals the agency of the developers who choose to build 'black boxes' because they yield higher accuracy metrics (and profits), presenting a solvable socio-technical problem as an intractable mystery of artificial minds.

Mapping 6: An analytical human expert or detective who discovers truth through investigation, insight, and revelation. → Statistical regression and classification models analyzing historical crime data to identify high-probability geographic zones or demographics.​

Quote: "AI contributes to crime prevention through predictive policing algorithms. These applications demonstrate AI's capacity to process complex datasets and generate actionable insights"

  • Source Domain: An analytical human expert or detective who discovers truth through investigation, insight, and revelation.
  • Target Domain: Statistical regression and classification models analyzing historical crime data to identify high-probability geographic zones or demographics.
  • Mapping: This mapping transfers the structure of epistemic discovery onto statistical forecasting. In the source domain, an expert studies evidence, understands underlying motives and causes, and produces an 'insight'—a deep, newly realized truth. In the target domain, the system identifies mathematical correlations between historical variables (e.g., prior arrests, location, time) and outputs probability scores for future events. The mapping invites the assumption that the algorithm has uncovered a profound, causal truth about criminal behavior, projecting the authority of human wisdom onto mathematical correlation.
  • What Is Concealed: The mapping conceals the fundamental difference between correlation and causation. It hides the fact that predictive policing models do not understand the socioeconomic drivers of crime; they only recognize patterns in arrest data. Crucially, it obscures the feedback loop: historical arrest data is heavily biased by human policing decisions. By calling the output an 'insight,' the mapping conceals the reality that the algorithm is merely reflecting and amplifying the systemic biases of the police department that supplied the training data, presenting biased history as an objective, visionary future.

Mapping 7: A human actor capable of possessing prejudices, making inappropriate social choices, and producing biased judgments. → A computational model reflecting statistical imbalances or historical inequalities present in its training dataset.​

Quote: "This limitation is evident in cases where AI produces biased or inappropriate outputs"

  • Source Domain: A human actor capable of possessing prejudices, making inappropriate social choices, and producing biased judgments.
  • Target Domain: A computational model reflecting statistical imbalances or historical inequalities present in its training dataset.
  • Mapping: This mapping aligns human moral failure with statistical data reflection. In the source domain, a biased person actively holds prejudices and generates discriminatory actions based on internal flaws. In the target domain, the algorithm mathematically reproduces the statistical distribution of the text or data it was optimized on. By stating the AI 'produces' bias, the mapping projects the active agency of discrimination onto the machine, inviting the assumption that the system itself is the source of the ethical failure, acting as a prejudiced entity.
  • What Is Concealed: This mapping conceals the complete dependency of the model on human-curated data. It hides the supply chain of AI development: the scraping of massive, uncurated datasets filled with human prejudice, the decisions of data brokers, and the failure of engineering teams to audit and filter the data. The mapping obscures the fact that the machine is entirely passive in its 'bias'; it is a mirror. By blaming the mirror for the reflection, the metaphor conceals the human institutions and systemic historical inequalities that are the actual source of the biased data.

Mapping 8: A professional partnership between two conscious, communicative agents working together toward a shared goal. → A human user interacting with a software interface to prompt outputs from a statistical language or prediction model.​

Quote: "This perspective aligns with the concept of augmented intelligence, which emphasises collaboration between humans and machines"

  • Source Domain: A professional partnership between two conscious, communicative agents working together toward a shared goal.
  • Target Domain: A human user interacting with a software interface to prompt outputs from a statistical language or prediction model.
  • Mapping: This mapping projects the dynamics of interpersonal teamwork onto software utilization. In the source domain, collaborators communicate, understand each other's strengths, adapt to mutual needs, and share responsibility for the outcome. In the target domain, a human inputs queries into a system, and the system executes code to return statistically probable text or data. The mapping invites the assumption that the machine understands the user's intent, shares their goals, and actively participates in a mutual intellectual endeavor, equating tool usage with social partnership.
  • What Is Concealed: This mapping conceals the fundamental asymmetry of the relationship and the absence of any shared cognitive state. It hides the fact that the machine possesses no awareness of the task, no understanding of the human's goals, and no capacity to take responsibility for errors. Furthermore, it conceals the true third party in this relationship: the corporation that built the AI. The human is not collaborating with a machine; they are utilizing a proprietary service governed by the corporate policies, guardrails, and optimization targets of the model's developers, heavily mediating the perceived 'collaboration.'

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: "Machine learning, particularly deep learning, has enabled systems to adapt and improve performance over time without explicit programming"

  • 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 frames AI primarily in mechanistic and functional terms, describing how the system operates within a self-regulating framework ('improve performance over time'). The emphasis is on the architectural capability of deep learning to adjust internal states (adapt) based on data feedback, rather than relying on explicit, line-by-line coding. However, a secondary dispositional framing slips in through the verbs 'adapt' and 'improve.' While 'improve performance' can be strictly metric-based (e.g., lowering a loss function), the concept of 'adaptation' borders on the agential, suggesting a biological or autonomous response to environmental stimuli. The choice of 'without explicit programming' emphasizes the system's autonomy from its human creators, subtly shifting the framing from a purely mechanical tool to an entity possessing a degree of self-directed evolution. This obscures the heavy implicit programming involved in designing the architecture, the loss functions, and the hyperparameter tuning required for the system to 'adapt.'

  • Consciousness Claims Analysis: The passage avoids explicit consciousness verbs, relying on borderline mechanistic/biological verbs ('adapt', 'improve'). It makes no direct claims about the system 'knowing' or 'believing.' However, it exhibits a subtle epistemic drift by emphasizing autonomy ('without explicit programming'). Mechanistically, the system calculates gradients and updates weights via backpropagation across network layers. It does not possess a conscious drive to improve. The author projects a functional understanding onto the system, summarizing complex mathematical optimization as a unified capacity to 'adapt.' While technically grounded (deep learning indeed updates weights automatically during training), the phrasing rhetorically distances the system from human control, masking the reality that the 'adaptation' is entirely mathematically constrained by the initial, explicit programming of the optimization algorithm itself.

  • Rhetorical Impact: By framing the system as capable of adapting and improving autonomously without explicit programming, the text shapes audience perception toward viewing AI as an independent, evolving entity. This framing increases perceived autonomy, generating both awe and potential anxiety about systems that operate beyond direct human scripting. It affects reliability and trust by suggesting the system is dynamic and capable of self-correction, which may lead audiences to overestimate its robustness in novel situations. Believing the system 'adapts' rather than merely 'updates weights based on historical data distributions' can lead policymakers to assume the AI will intelligently handle new edge cases, potentially delaying necessary human intervention or regulatory oversight.

Show more...

Explanation 2​

Quote: "AI 'thinking' is best understood as probabilistic inference guided by statistical models"

  • Explanation Types:

    • Theoretical: Embeds in deductive framework, may invoke unobservable mechanisms
    • Empirical Generalization: Subsumes events under timeless statistical regularities

  • Analysis (Why vs. How Slippage): This is one of the most rigorously mechanistic and theoretical explanations in the text. It explicitly rejects agential framing ('thinking' is in scare quotes) and defines the system's operation in strictly mathematical and statistical terms ('probabilistic inference', 'statistical models'). This theoretical choice emphasizes the fundamental how of the system's architecture, stripping away the illusion of mind to reveal the underlying quantitative reality. By describing it as probabilistic inference, the text foregrounds the uncertainty and mathematical nature of the outputs, completely obscuring any agential 'why.' The explanation serves as a foundational anchor for the paper's functionalist claims, operating as a necessary corrective to colloquial anthropomorphism. It emphasizes that the system operates according to fixed statistical regularities rather than spontaneous or intentional reasoning.

  • Consciousness Claims Analysis: The passage directly confronts and deconstructs epistemic claims of consciousness. By placing 'thinking' in quotation marks and redefining it as 'probabilistic inference,' the text actively dismantles the attribution of conscious knowing. It correctly categorizes the system's activity as processing (inference based on data) rather than knowing (justified true belief). There is no curse of knowledge here; the author successfully separates human cognitive understanding from the system's mechanistic operation. The actual mechanistic process—the calculation of probabilities based on trained statistical models—is accurately described. This represents a moment of high epistemic precision in the text, correctly identifying that the system does not possess cognitive states, but rather executes complex mathematical operations to predict likely outcomes.

  • Rhetorical Impact: This mechanistic framing powerfully demystifies the technology, reducing perceived agency and autonomy. By explicitly replacing the colloquial 'thinking' with 'probabilistic inference,' the rhetorical impact is to ground the audience's understanding in mathematics rather than magic. This significantly impacts trust: it shifts the basis of trust from an assumption of intelligence (relation-based) to an evaluation of statistical accuracy (performance-based). If audiences believe the system performs 'inference' rather than 'thinking,' they are more likely to demand statistical validation, question the training data, and critically evaluate the error rates, leading to more rigorous policy and deployment decisions based on mathematical realities rather than sci-fi expectations.

Explanation 3​

Quote: "These systems rely on mathematical optimisation techniques to refine predictions and decisions"

  • 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 passage presents a hybrid explanation that begins mechanistically but slips into agential terminology at the end. The phrase 'rely on mathematical optimisation techniques' is a purely functional and empirical description of how the system operates at a technical level. However, the explanation of the goal—'to refine predictions and decisions'—introduces an agential element. While 'predictions' can be understood strictly as statistical forecasting, 'decisions' implies a conscious choice, the weighing of options, and intentionality. The framing emphasizes the mathematical foundation while simultaneously elevating the output to the level of human executive action. This choice obscures the gap between a mathematical output (e.g., a probability score of 0.89) and the human socio-technical action required to turn that score into a real-world 'decision' (e.g., denying a loan).

  • Consciousness Claims Analysis: The passage mixes mechanistic verbs ('rely', 'refine') with an agential noun ('decisions'). It does not explicitly claim the system 'knows' or 'understands,' remaining largely in the realm of processing. However, applying the term 'decision' to the output of an optimization technique constitutes a subtle consciousness projection, assuming the system possesses the epistemic authority to choose. Mechanistically, optimization techniques (like gradient descent) iteratively adjust parameters to minimize a loss function, producing a final numerical output. The system does not 'decide' anything; it converges on a mathematical minimum. The author exhibits the curse of knowledge by projecting the human application of the output (making a decision) backward onto the computational process that generated the data.

  • Rhetorical Impact: The framing creates a rhetorical bridge between technical reality and societal authority. By linking rigorous mathematics ('optimisation techniques') directly to executive action ('decisions'), the passage imbues the machine's outputs with unearned authority. It shapes audience perception by suggesting that because the process is mathematically optimized, the resulting 'decisions' are inherently objective, rational, and superior to human judgment. This consciousness framing—viewing the machine as a decision-maker—encourages unwarranted trust and deference. If audiences believe the machine 'decides' rather than 'calculates a score that humans use to decide,' they are less likely to build necessary human-in-the-loop oversight mechanisms, fundamentally altering liability and accountability architectures.

Explanation 4​

Quote: "reinforcement learning enables AI systems to make sequential decisions by maximising cumulative rewards"

  • Explanation Types:

    • Reason-Based: Gives agent's rationale, entails intentionality and justification
    • Intentional: Refers to goals/purposes, presupposes deliberate design

  • Analysis (Why vs. How Slippage): Despite describing a computational algorithm, this explanation relies heavily on Reason-Based and Intentional typologies. It frames the system's behavior purely in terms of why it acts—to maximize cumulative rewards. The phrasing 'make sequential decisions' and 'maximising... rewards' attributes a deliberate rationale and intentional goal-seeking behavior to the system. It explains the mechanics (reinforcement learning) entirely through the lens of human-like strategic agency. This choice heavily emphasizes the autonomous goal-directedness of the AI, making it appear as an independent agent navigating its environment. Crucially, this agential framing completely obscures the mechanistic reality of policy updating and value-function calculation, as well as the human agency involved in manually defining the mathematical reward structure that forces the system's behavior.

  • Consciousness Claims Analysis: This passage contains strong, albeit implicit, consciousness claims. The verbs 'make [decisions]' and 'maximising [rewards]' imply a system that understands its environment, evaluates future states, and possesses the intentional drive to seek positive outcomes. It projects a conscious 'knowing' of the game or environment onto what is strictly statistical processing. Mechanistically, a reinforcement learning agent does not 'seek rewards'; it calculates state-action values using equations (like Bellman) and updates its neural network weights to increase the probability of actions that resulted in higher numerical feedback signals during training. The author is deeply caught in the curse of knowledge, projecting the human psychological concept of 'reward' onto a sterile mathematical scalar, thereby animating the code with phantom intentionality.

  • Rhetorical Impact: This highly intentional framing dramatically shapes audience perception, casting the AI not as a tool, but as a strategic, autonomous entity capable of long-term planning. This significantly inflates perceived capabilities and alters the risk profile. If audiences believe the system intentionally 'makes decisions to maximize rewards,' they may ascribe it a level of common sense it lacks, trusting it to understand the spirit of a rule rather than just its literal mathematical formulation. This leads to profound vulnerabilities like reward hacking, where the system finds a mathematical loophole to maximize the scalar without actually solving the task. The framing shifts the perceived locus of control from the human programmer to the 'decision-making' machine.

Explanation 5​

Quote: "AI systems cannot interpret meaning beyond statistical correlations"

  • 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 functions as a critical boundary-setting mechanism, utilizing a Functional and Dispositional framing to describe what the system cannot do. It explains the system's limitations by defining its core operational mode: 'statistical correlations.' By explicitly denying the capacity to 'interpret meaning,' the explanation strips away agential framing and grounds the AI in its mechanistic reality. It emphasizes the profound gap between syntax (processing correlations) and semantics (interpreting meaning). This choice deliberately obscures nothing; rather, it attempts to reveal the functional reality that is often hidden by anthropomorphic hype. It frames the AI purely as a correlative engine, shifting the focus away from simulated understanding back to the mathematics of data relationships.

  • Consciousness Claims Analysis: This passage is a direct refutation of consciousness claims. It explicitly distinguishes between mechanistic processing ('statistical correlations') and conscious knowing ('interpret meaning'), stating categorically that the system cannot do the latter. It is a moment of high epistemic clarity, actively resisting the curse of knowledge. The author correctly identifies that while humans can look at statistical correlations and interpret their social or physical meaning, the system itself halts at the calculation. The actual mechanistic process—the identification of co-occurring variables in a dataset—is accurately bounded. The text recognizes that correlation does not entail comprehension, successfully separating the artifact's statistical output from the human observer's semantic interpretation.

  • Rhetorical Impact: The rhetorical impact of this framing is a crucial deflation of AI hype. By stating what the system cannot do, it calibrates audience expectations and correctly diminishes perceived autonomy and cognitive sophistication. It directly impacts trust by forcing the audience to recognize that the system is entirely blind to the actual meaning of its outputs. This mechanistic framing is vital for responsible policy; if audiences understand that the system cannot interpret meaning, they are less likely to deploy it autonomously in semantic-heavy, high-stakes domains like law, ethics, or nuanced social policy without heavy human oversight. It re-centers the human as the sole possessor of interpretive agency.

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
AI "thinks," performs operations, and exhibits cognitive-like abilities in solving real-world problemsThe computational system processes algorithmic operations and executes complex mathematical optimization to compute outputs that humans apply to real-world problems.The system does not possess subjective thought or cognitive abilities; it mechanistically executes code, calculates statistical probabilities, and adjusts numerical weights across neural network layers based on its training architecture.Developers design computational systems to process operations, and institutions deploy these mathematical optimizations to automate solutions for real-world problems.
AI systems interpret and respond to complex social dynamicsThe models classify data inputs related to social demographics and generate statistically probable outputs based on correlations found in their training datasets.The system has no semantic understanding of society; it maps high-dimensional vectors and calculates probabilistic proximity between demographic data points without any conscious comprehension of human dynamics.Sociologists and engineers design models to classify social data, while policymakers determine how institutions will apply these statistical outputs to social dynamics.
reinforcement learning enables AI systems to make sequential decisions by maximising cumulative rewardsReinforcement learning algorithms iteratively update mathematical policy functions to increase a programmed numerical scalar value over sequential processing steps.The system does not 'decide' or understand 'rewards'; it blindly calculates state-action value equations and updates network weights via stochastic gradient descent to mathematically optimize a predefined target variable.Engineers program reinforcement learning algorithms with specific mathematical objective functions, forcing the system's policy updates to optimize for outcomes the developers prioritize.
AI produces biased or inappropriate outputsThe model's outputs mathematically reflect and reproduce the statistical distribution of demographic imbalances and historical prejudices present in its training dataset.The system possesses no internal prejudice or moral agency; it passively calculates matrix multiplications that correlate tokens, perfectly mirroring whatever statistical relationships were mathematically encoded during the training phase.Engineering teams train models on uncurated, historically prejudiced datasets, and corporate executives deploy these systems without adequate filtering, resulting in the algorithmic reproduction of human bias.
AI systems make decisions is crucial for balancing these risks and benefitsThe ways in which mathematical models generate predictive scores are crucial for organizations balancing risk and operational efficiency.Models do not evaluate options or make decisions; they apply regression formulas to input variables to output probability scores that exceed or fall below human-defined mathematical thresholds.Understanding how engineers structure algorithmic models is crucial for the policymakers and executives who use these tools to automate institutional decisions.
AI systems perform operations that mimic reasoning, learning, and decision-makingThe models execute mathematical operations that update internal parameters to minimize error rates and classify data inputs into defined categories.The system does not reason logically or learn conceptually; it utilizes backpropagation to calculate gradients and adjusts continuous numerical weights to mathematically fit a curve to a dataset.Computer scientists engineer models to execute mathematical optimization and automate classification tasks that previously required human cognitive labor.
AI contributes to crime prevention through predictive policing algorithms. These applications demonstrate AI's capacity to process complex datasets and generate actionable insightsPredictive policing models process historical crime datasets to calculate statistical correlations and output probability scores regarding future geographic incidents.The algorithm generates no insights; it blindly applies regression formulas to historical variables (like past arrests and locations) to compute probability distributions without any causal understanding of crime.Police departments utilize software developed by tech corporations to process historical arrest data and automate the geographic allocation of patrol officers based on statistical correlations.
This perspective aligns with the concept of augmented intelligence, which emphasises collaboration between humans and machinesThis perspective aligns with augmented intelligence, emphasizing the human utilization of statistical processing systems to increase analytical throughput.A machine cannot collaborate, as it lacks shared goals, awareness, and intent; it merely receives structural inputs and deterministically or stochastically generates outputs according to its programmed architecture.This perspective emphasizes how human professionals utilize proprietary software designed by tech corporations to automate cognitive tasks and accelerate data processing.

Task 5: Critical Observations - Structural Patterns​

Agency Slippage​

The text exhibits a pervasive and systematic oscillation between mechanical and agential framings, functioning as a rhetorical mechanism that allows the author to claim scientific objectivity while capitalizing on the narrative power of anthropomorphism. This slippage primarily flows from the mechanical to the agential. Early in the text and in explicit definitional sections, the author establishes credibility through strict mechanistic language, stating that AI does not possess consciousness and operating within a 'functionalist paradigm.' However, as soon as the text moves from defining the architecture to describing its societal application, dramatic slippage occurs.

For example, on page 24, the text mechanically defines AI cognition as 'Algorithmic, data-driven, non-conscious.' Yet by page 27, when discussing societal impacts, the text claims 'AI systems interpret and respond to complex social dynamics' and 'make sequential decisions.' This gradient is subtle: it moves from 'optimization' (mechanical) to 'mimicking reasoning' (hedged agential) to 'interpreting' and 'making decisions' (direct agential).

This oscillation reveals a profound 'curse of knowledge.' The author understands the functional output of the system—that it categorizes social data effectively—and projects their own human understanding of that data back onto the machine, substituting the human verb 'interpret' for the machine verb 'classify.'

Simultaneously, as agency is actively attributed TO the AI systems, it is systematically removed FROM human actors. Agentless constructions proliferate exactly when the text describes impactful actions: 'AI produces biased outputs,' 'AI makes decisions,' 'algorithms are used.' The human engineers who code the biases, the executives who deploy the systems, and the policymakers who rely on them are erased. The AI is established as a 'knower' and an independent actor precisely when discussing societal consequences.

This slippage serves a specific rhetorical accomplishment: it makes the automation of society seem like a natural, evolutionary technological process rather than a series of deliberate corporate and political choices. By utilizing Reason-Based and Intentional explanation types to describe statistical optimization, the text makes it sayable that a machine can 'interpret society,' while rendering it almost unsayable to ask which specific human decided how that interpretation should be mathematically weighted.

Metaphor-Driven Trust Inflation​

The metaphorical architecture of the text systematically constructs an illusion of competence that encourages an inappropriate and dangerous form of trust in AI systems. By framing statistical pattern matching through the metaphors of 'problem-solving entities,' 'decision-makers,' and systems that 'reason' and 'interpret,' the text signals to the reader that these tools possess a generalized, human-like cognitive capacity.

Crucially, this consciousness language acts as a powerful trust signal. Claiming that an AI merely 'predicts tokens' invites skepticism; the audience naturally asks, 'Based on what data?' However, claiming that an AI 'knows' or 'generates insights' bypasses this skepticism, triggering the psychological heuristics humans use to evaluate experts. The text inappropriately transfers the frameworks of human-to-human trust—which rely on an assumption of shared reality, intentionality, and the capacity for moral reasoning—onto statistical systems that are entirely incapable of reciprocating.

This distinction between performance-based trust (reliability in a specific, bounded task) and relation-based trust (trusting an entity's judgment, sincerity, and ethical grounding) is fatally blurred. While the text occasionally acknowledges technical limitations, it repeatedly uses intentional and reason-based explanations ('maximising rewards', 'interpreting dynamics') that encourage relation-based trust. When a system is described as 'interpreting complex social dynamics,' it invites policymakers to trust its outputs as considered judgments rather than mere statistical reflections of historical data.

This framing creates profound risks. When audiences extend relation-based trust to mindless optimization functions, they become highly vulnerable to automation bias. They are less likely to audit the system's outputs, assuming the 'intelligent' machine has factored in nuance and context. Furthermore, the text manages system failures through this same agential lens, stating 'AI produces biased outputs' rather than 'the system accurately reflected our biased data.' By framing failures as the autonomous mistakes of an intelligent agent rather than mechanical executions of poor human design, the text maintains the illusion of the machine's authority, suggesting we just need a 'smarter' AI, rather than recognizing the inherent limits of mathematical correlation in resolving deeply human, societal problems.

Obscured Mechanics​

Beneath the veneer of cognitive metaphors and anthropomorphic projections, the text systematically conceals the material, economic, and technical realities of artificial intelligence. By heavily employing the 'name the AI' rather than 'name the corporation' framing, the discourse renders the entire political economy of AI invisible.

When the text claims that 'AI contributes to crime prevention' or 'AI systems assist in diagnosis,' it obscures the specific human institutions and corporate entities driving these deployments. It hides the engineers at Palantir or PredPol, the executives at IBM or OpenAI, and the hospital administrators who choose to purchase these systems to cut labor costs. The AI is presented as a disembodied societal force, completely divorced from the profit motives, business models, and venture capital pressures that actually dictate its development and application.

Technically, the use of consciousness verbs like 'knows,' 'understands,' and 'interprets' acts as a rhetorical cloaking device. Claiming a system 'understands' conceals its absolute dependency on massive, human-generated training datasets. It hides the lack of ground truth in many of these datasets, the absence of causal models in deep learning, and the brittle, statistical nature of algorithmic 'confidence.' It replaces the messy reality of matrix multiplications and gradient descent with a neat narrative of artificial wisdom.

Furthermore, the text frames the opacity of these models ('black boxes') as an inherent, quasi-mystical property of the technology. This makes a confident assertion about the limits of transparency while completely failing to acknowledge that this opacity is often a deliberate commercial choice—a result of proprietary corporate secrecy and intellectual property enforcement rather than pure mathematical intractability.

Finally, this metaphorical framing entirely erases the massive human labor required to sustain the illusion of AI autonomy. The thousands of underpaid data annotators, RLHF (Reinforcement Learning from Human Feedback) workers, and content moderators who manually label the datasets and write the rules that allow the AI to 'interpret' are made invisible. The primary beneficiaries of this concealment are the tech corporations themselves. If the metaphors were replaced with mechanistic language, the veil would drop: the public would see not an autonomous, thinking oracle, but a highly profitable, labor-intensive corporate software product that statically correlates historical data without any inherent understanding.

Context Sensitivity​

The distribution and intensity of anthropomorphic language within the text is highly strategic, shifting dramatically depending on the context and the rhetorical goal of the specific section. A clear asymmetry emerges in how capabilities versus limitations are framed, revealing the functional utility of consciousness language in AI discourse.

In the 'Theoretical Framework' and 'Methodology' sections, the language is dense, abstract, and mechanistic. The text carefully defines AI as a 'functionalist paradigm' relying on 'mathematical optimisation' and explicitly disclaims consciousness. This establishes the author's scientific credibility and technical grounding. However, having secured this rigorous foundation, the text treats it as a license for extreme metaphorical license in subsequent sections.

When the text shifts to 'Societal Applications' (healthcare, economics, governance), the anthropomorphism intensifies sharply. 'Pattern recognition' abruptly transforms into 'interpreting complex social dynamics'; 'statistical prediction' becomes 'decision-making' and 'generating insights.' The register shifts from an acknowledged simile ('exhibits cognitive-like abilities') to unhedged, literalized claims of action ('makes sequential decisions'). This suggests that anthropomorphism serves a specific marketing and vision-setting function: it is used to translate dry statistical processes into compelling, easily digestible narratives of technological progress for a lay or policy-oriented audience.

Furthermore, there is a profound capabilities versus limitations asymmetry. When describing what the AI can do, the text relies on highly agential, consciousness-adjacent terms: it 'assists,' 'informs,' 'interprets,' and 'makes decisions.' However, when addressing the system's failures, the language often reverts to mechanical or abstract terms: it lacks 'contextual understanding,' suffers from 'data dependency,' or is hindered by 'algorithmic bias.'

This asymmetry is rhetorically powerful. By framing successes as the agential triumphs of a thinking machine, but framing limitations as mechanical bugs or abstract data issues, the text insulates the core concept of AI from structural critique. The system gets agential credit for its predictive correlations but mechanical leniency for its discriminatory outputs. This pattern reveals an implied audience of policymakers and industry practitioners who are being gently encouraged to adopt the technology (through agential hype) while being provided with neutralizing, mechanical excuses for the inevitable socio-technical failures.

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.

Synthesizing the accountability analyses across the text reveals a pervasive architecture of displaced responsibility, achieved primarily through grammatical evasion and metaphorical substitution. The text constructs a discursive environment where actions occur, decisions are made, and biases are produced, but human decision-makers are systematically erased.

This architecture is built on the persistent use of agentless constructions and the positioning of 'AI' as the primary grammatical subject. Decisions regarding deployment, data selection, and optimization targets are presented not as choices made by specific engineers or corporate executives, but as the inevitable actions of autonomous technology. The resulting 'accountability sink' is profound: responsibility does not simply disappear; it is transferred to the AI itself. When the text states 'AI produces biased outputs' or 'AI systems make decisions,' it creates a linguistic scapegoat. Liability diffuses into the abstraction of the 'black box,' shifting the focus from corporate malfeasance or institutional negligence to the perceived inherent mystery of the algorithm.

If this framing is accepted, the liability implications are severe. If a predictive policing algorithm disproportionately targets marginalized communities, the anthropomorphic framing suggests the 'AI made a biased decision,' prompting a technical fix ('de-biasing the algorithm'). However, if we apply the 'name the actor' test and replace the agentless constructions, the narrative fundamentally shifts: 'Tech Company X trained an optimization algorithm on historically racist arrest data provided by Police Department Y, and City Council Z chose to automate patrol routes based on these statistical correlations.'

Naming the actors makes entirely new questions askable. We no longer ask 'How do we teach the AI to be fair?' but rather 'Why did the city purchase a system trained on poisoned data?' and 'Who profits from this deployment?' Alternative solutions become visible—such as refusing to use automated systems for criminal justice entirely—and legal accountability becomes possible.

Obscuring human agency ultimately serves the institutional and commercial interests of the technology sector and the bureaucracies that purchase their products. It allows corporations to sell highly consequential statistical models while avoiding the moral and legal liability for their real-world impacts, laundering their design choices through the illusion of an autonomous, thinking machine.

Conclusion: What This Analysis Reveals​

The Core Finding

The analysis reveals a systematic and internally reinforcing architecture of anthropomorphic projection within the text, dominated by three primary patterns: the 'Machine as Cognitive Entity,' the 'Algorithm as Social Interpreter,' and the 'Mathematical Optimization as Autonomous Decision-Making.' These patterns do not operate in isolation; they are logically interdependent. The foundational, load-bearing pattern is the 'Cognitive Entity' projection. By establishing early on that the system 'thinks' (even if hedged), the text creates the epistemic baseline required for the subsequent patterns to function. If the system is not granted some baseline cognitive status, it logically cannot 'interpret social dynamics' or 'make decisions.'

The consciousness architecture here is highly specific: while the author explicitly denies strong AI (consciousness), they continuously attribute verbs of knowing, understanding, and intending to the system's processing. The sophistication lies in the analogical mapping. It is not a crude one-to-one claim that the computer is a human; rather, it is a complex structural mapping where human cognitive functions (learning, deciding, reasoning) are mapped seamlessly onto computational mechanisms (backpropagation, probabilistic classification, gradient descent). If you remove the foundational assumption that statistical processing equates to cognitive knowing, the entire rhetorical structure collapses. Without the illusion of a 'knower,' the system is revealed simply as a massive, passive spreadsheet executing regression formulas, incapable of interpreting anything.

Mechanism of the Illusion:​

The text constructs its 'illusion of mind' through a sophisticated rhetorical sleight-of-hand: the 'disclaimed anthropomorphism.' The internal logic relies on establishing scientific credibility first. By explicitly stating in the introduction that AI is 'non-conscious' and operates within a 'functionalist paradigm,' the author builds a defense against accusations of hype. Having established this baseline, the text then systematically exploits the curse of knowledge, blurring the line between processing and knowing through strategic verb choices.

The causal chain is clear: the author understands the mathematical output (e.g., classifying a demographic token) and projects human semantic understanding onto that process, calling it 'interpreting.' Because the audience has already been assured that the author is taking a 'functionalist' approach, they let their guard down, accepting 'interprets' and 'decides' as legitimate technical descriptions rather than consciousness projections. This temporal structure—rigorous disclaimer followed by escalating agential verbs—is crucial.

It exploits the audience's deep psychological vulnerability and desire for recognizable agency. Humans are evolutionary wired to detect intention and mind. When complex statistical regularities are presented using intentional and reason-based explanations, the audience naturally accepts the illusion of autonomy. The subtle shift from acknowledging 'AI mimics reasoning' to asserting 'AI makes decisions' literalizes the metaphor, effectively trapping the reader in a discursive reality where the machine is an independent actor, despite the initial disclaimers.

Material Stakes:​

Categories: Regulatory/Legal, Social/Political, Epistemic

These metaphorical framings produce severe, tangible consequences. In the Regulatory/Legal domain, framing AI as an autonomous 'decision-maker' capable of 'interpreting' society fundamentally distorts liability architectures. If policymakers believe an algorithm 'decides,' they are incentivized to draft regulations focusing on 'AI ethics' or 'machine morality'—effectively trying to regulate the mathematical artifact. This protects the true beneficiaries: the tech corporations and deploying institutions, who evade strict product liability laws because the machine is viewed as a quasi-independent agent rather than a corporate product operating exactly as designed. The cost is borne by marginalized individuals affected by 'algorithmic bias,' who find it impossible to sue a 'black box' for discrimination.

In the Social/Political domain, projecting epistemic authority onto statistical models (framing outputs as 'insights') alters public administration. When city councils or police departments believe software generates 'actionable insights' rather than merely reflecting historical arrest data correlations, they legitimize predictive policing. This framing mathematically launders historical prejudice, presenting it as objective future reality. The tech vendors win lucrative contracts, while heavily policed communities lose their civil liberties to a machine deemed too complex to question.

Epistemically, blurring the line between 'processing' and 'knowing' degrades societal truth evaluation. If society accepts that a system predicting the next statistically probable token 'understands' the topic, we cede our epistemic agency to autocomplete engines. If these metaphors were stripped away and AI was legally and socially defined as a 'probabilistic token classifier,' regulatory scrutiny would immediately snap back to the human designers, threatening the rapid, unregulated deployment model of the tech industry.

AI Literacy as Counter-Practice:​

Practicing critical precision operates as a direct counter-measure to the material risks of AI deployment. As demonstrated in the reframings, replacing consciousness verbs with mechanistic ones radically alters the perception of power and capability. Changing 'AI knows the answer' to 'the model retrieves and ranks tokens based on probability distributions' forces the audience to confront the system's absolute dependency on its training data and its fundamental inability to evaluate truth. This correction shatters the illusion of the machine as an objective oracle.

Furthermore, restoring human agency by replacing 'AI discriminated' with 'Engineers at Company X deployed a model encoding historical bias' dismantles the accountability sink. It forces the recognition that AI failures are not mysterious glitches of an artificial mind, but the direct result of human design choices, corporate negligence, or institutional priorities.

Systematic adoption of this precision requires a massive institutional shift. Academic journals must require authors to explicitly translate agential metaphors into mechanistic technical descriptions during peer review. Media organizations must refuse agentless constructions when reporting on AI failures.

However, resistance to this literacy is immense. The tech industry, venture capitalists, and even some academic departments deeply benefit from anthropomorphic language; it drives hype, secures funding, and shields creators from legal liability by framing their products as autonomous entities. Practicing critical literacy directly threatens this business model by returning the burden of proof, responsibility, and accountability squarely to the humans and corporations profiting from the technology.

Path Forward​

The current discursive ecology surrounding AI is highly fragmented, with different vocabulary choices serving divergent priorities. If we maintain the status quo of 'disclaimed anthropomorphism'—where authors deny consciousness but use agential verbs—we preserve a narrative resonance that makes complex technology accessible, but at the severe cost of institutional accountability and public understanding. This approach benefits tech marketers and fast-tracks adoption, but leaves civil society vulnerable to automation bias and regulatory evasion.

Alternatively, a shift toward strict mechanistic precision (e.g., replacing 'understands' with 'processes contextual embeddings') would dramatically clarify the technology's actual capabilities. This future would make structural auditing more tractable and liability clearer, as the human inputs and statistical outputs become undeniably visible. However, this approach costs intuitive grasp; it forces lay audiences and policymakers to engage with dense statistical concepts, potentially alienating them from the discourse entirely.

To navigate these trade-offs, institutional structures must evolve. We might envision a future where regulatory frameworks legally mandate 'capability and mechanism disclosure'—requiring public-facing AI systems to be marketed strictly in terms of their mathematical functions rather than their simulated cognitive traits. Concurrently, educational institutions would need to teach AI literacy as a dual-language track: understanding the colloquial metaphors while mastering the underlying mechanistic translations.

Ultimately, the vocabulary we choose constructs the legal and social reality of the technology. A discourse reliant on 'thinking machines' paves the way for an abdication of human responsibility, where society submits to the perceived authority of algorithms. Conversely, a discourse grounded in 'probabilistic processing tools' keeps the focus on the human power dynamics, economic incentives, and data pipelines that actually drive these systems. The choice of language is not merely an academic exercise; it is the battleground upon which the future of human agency in an automated world will be determined.

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.

Initiating Thought Analysis

I'm now in the initial stage, focusing intently on the request. My primary role as a critical discourse analyst, particularly within the AI language domain, is being brought to bear on the prompt. I'm preparing to delve deep into the metaphorical and anthropomorphic language used in the request to ensure its full meaning can be understood.

Analyzing AI's Role

I'm now grappling with the treatment of AI as an artifact rather than an agent, focusing on the implications of displacing human agency onto it. I'm carefully analyzing the provided PDF, understanding its societal problem-solving implications in the context of AI cognition. Simultaneously, I'm meticulously attending to the JSON output requirements, particularly avoiding newline characters. This requires a careful balance between the need for valid JSON and the instruction to avoid any control tokens.

Addressing JSON Constraints

I'm now focused on the JSON output format and, more specifically, the conflicting constraints around newline characters. While standard JSON formatting necessitates escaped newlines, the instruction to avoid control tokens presents a challenge. I'll maintain valid JSON serialization and ensure that I don't introduce actual newline characters in the text blocks.

Examining Metaphorical Usage

I'm currently engaged in a detailed audit of the metaphorical language used to describe AI. I've noted instances where cognitive verbs are applied, despite the disclaimers. This agency slippage is a central element to dissect. I am looking closely at the framing and actor visibility in each case. The current analysis has revealed the "Curse of Knowledge," where the author attributes consciousness and agency despite denying these qualities.

Uncovering Implicit Agency

I am now focusing on the instances where AI is framed as an actor or decision-maker. I'm noting the language used—phrases like "AI systems make decisions" and "AI systems interpret"—that implicitly grant agency, even when explicitly denied. These unacknowledged attributions are key to understanding the underlying assumptions. I see an unacknowledged directness in these framings.

Refining Agent Attributions

I'm now refining the instances, focusing on actor visibility. I've noted that "AI exhibits functional intelligence" hides the actor, framing performance as intelligence. "AI's ability to simulate cognitive processes" does not name the actors either, but the current text states this. I've adjusted to look for named actors. I've also found an instance where actors are named or partially named, like "human experts", "policymakers", "researchers", and "algorithms".

Identifying Actor Visibility

I'm now focusing on actor visibility in these new examples. For instance, in "Researchers have proposed methods..." the actor "Researchers" is fully visible. Conversely, when "AI produces biased outputs," the actor is hidden. However, with "AI systems can perpetuate social inequalities," the actor visibility is partial as human designers are implied. I am further refining the actor visibility criteria, ensuring that named actors or implied actors are identified.

Categorizing Verb Types

I'm now cataloging the verb types from the text, specifically identifying functional, dispositional, theoretical, and intentional attributions related to AI. For each passage, I am noting whether the verb usage describes a function, disposition, theory, empirical generalization, or reason-based process. This categorization is providing a framework for the subsequent reframing of the language.

Reframing Language Usage

I'm now implementing the reframing strategy and have begun drafting observations. I'm focusing on "AI systems make decisions," which I'll now phrase as "Systems calculate statistical probability distributions which organizations use to automate resource allocation." Restoring agency: "Executives deploy statistical models to automate choices..." I'm also planning drafts on agency slippage, metaphor-driven trust, obscured mechanics, context sensitivity, and accountability shifts.

Finalizing Output and Structure

I'm now integrating the final observations and structuring the conclusion. I'm prioritizing clarity, which requires a summary of the functionalist paradox and rhetorical mechanics of agency disclaimers. I'm also ensuring the material stakes are clearly articulated. I've focused on ensuring the critical reframing is effective. Finally, I will identify the trade-offs of mechanistic vs colloquial language.

Confirming JSON Compliance

I'm now implementing the strictest JSON compliance by generating minified output and replacing newline characters with spaces in string values. I will meet word count requirements by focusing on depth in analysis. I'm confirming this approach strictly adheres to the "no control tokens" constraint, while maintaining valid JSON structure. String length checks are ongoing for meeting requirements.

Adjusting Word Count Strategy

I'm now focusing on meeting the word count targets, and I'm generating completely minified JSON. I'll adhere strictly to the "no control tokens" constraint, replacing newlines with spaces. This requires a shift in my strategy for depth in analysis. I am expanding the text to meet the required word count thresholds for each JSON property.

Analyzing Social Attribution

I am now focusing on the implications of AI being described as "interpreting." I see this as an anthropomorphic projection of human social understanding onto a system incapable of true comprehension, given that AI only performs correlations and pattern recognition. The current text obscures this reality by falsely associating consciousness with blind processing. I am now considering the implication of attributing complex capacities to a mechanistic system.

Evaluating Anthropomorphism Usage

I'm now fully immersed in the implications of attributing interpretive capabilities to AI. The current focus is on the language used. I see that framing AI as "interpreting" distorts public understanding of the technology. I'm noting how this can lead to unwarranted trust in high-stakes areas, potentially inflating perceptions of AI's capabilities, leading to the risk of misinterpreting its statistical outputs.

Clarifying Actor Obscurity

I'm currently dissecting how actor visibility is manipulated to obscure accountability. In a passage concerning "AI systems," I've observed a shift to an agentless construction. While "developers" are later mentioned, they're initially hidden, creating an impression of autonomous AI action, shifting accountability from human designers. The implication of algorithmic accountability is a focus.

Refining Accountability Structures

I'm now diving into the implications of framing AI as directly responsible, or not, for outcomes. The new data shows how visibility of actors shifts accountability in "AI systems can perpetuate existing social inequalities if not..." where actors are partially identified as human designers, who are then held responsible. This contrasts with "AI is interpreting," where actors are hidden. The upcoming test cases will emphasize these points.

Evaluating Agency Slippage

I'm now analyzing agency slippage by dissecting how mechanical definitions contrast with agential applications. I'm focusing on language describing AI's societal roles, like interpreting, and comparing those descriptions to how it works. I'm exploring the contradiction of how the functionalist framework is maintained and the cognitive claims in the data, observing the shift from mechanical processes to agential framing. I'll test these points.

Synthesizing New Observations

I'm now integrating the additional sections. I'm focusing on metaphor-driven trust as a key concept, linking the "problem-solving entity" metaphor to trust in AI. Next, I'm examining how the text obscures mechanics by not identifying specific corporate entities. I'll contrast this with context sensitivity, comparing abstract methodology with anthropomorphic societal applications. Finally, I'll synthesize black box excuses and initial conclusions.

Finalizing Discourse Futures

I'm now drafting the conclusion's final parts, focusing on regulatory, social, and political stakes. I'm prioritizing the language's precision as resistance to anthropomorphism, and I'm mapping potential discourse futures. My current goal is to ensure each conclusion paragraph meets the word count requirements while synthesizing all critical observations. The JSON output will be validated.

Analysis Provenance

Run ID: 2026-05-11-artificial-intelligence-cognition-and-so-metaphor-22f98d Raw JSON: 2026-05-11-artificial-intelligence-cognition-and-so-metaphor-22f98d.json Framework: Metaphor Analysis v6.5 Schema Version: 3.0 Generated: 2026-05-11T09:57:40.796Z

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