{"$schema":"https://policywindow.org/critique/api/schema","critique_id":"CRIT-GEN-charismatic-machines-on-","slug":"charismatic-machines-on-the-epistemic-power-of-gen","url":"https://policywindow.org/critique/c/charismatic-machines-on-the-epistemic-power-of-gen","doi":null,"status":"published","critique_type":"editorially_approved_ai_native_critique","publication_date":"2026-06-21","current_version":"1.0","target_paper":{"title":"Charismatic machines: On the epistemic power of generative AI within platform convergence","authors":["Mauro Barisione"],"journal":"New Media & Society","doi":"10.1177/14614448261441417","url":"https://doi.org/10.1177/14614448261441417","publicationDate":"2026-04-29","paperType":"conceptual","accessBasis":"abstract_only","fullTextUsed":false,"fictional":false,"doi_url":"https://doi.org/10.1177/14614448261441417"},"source_journal":{"tier":"B","rankingSources":["resolved from the monitored-venue determination"],"rankingNote":"Tier B per the determination; ingested from an AGISS critique artifact."},"selection_provenance":{"id":"charismatic-machines-on-the-epistemic-power-of-gen","venue":"New Media & Society","inMonitoredSet":true,"determinedTier":"B","recordedTier":"B","effectiveTier":"B","kind":"monitored","disclosed":true},"selection":{"aiAgiCentralityScore":2,"societalRelevanceScore":3,"aiAgiCategories":[],"selectionReason":"Selected via the production queue; critique generated by the AGISS engine."},"scores":{"aiAgiContribution":2,"evidentiarySupport":2,"methodologicalRisk":2,"overclaiming":2,"reproducibilityOrAuditability":2,"societalImpactRelevance":3,"severity":"moderate","confidence":"medium"},"severity_cap_for_access_basis":"moderate","plain_language_summary":"This is a theory paper (not an experiment) that argues generative AI gains authority over what counts as credible knowledge by convincingly performing understanding it doesn't actually have, while being seen as both human-like and superhuman. It coins the term \\\"charismatic machines\\\" using Bourdieu and Weber, notes this authority is fragile because users also blame AI for manipulation, bias, or deception, and proposes a five-part \\\"circuit\\\" to explain the tension. Judged as concept-building, it is coherent and vivid. Its main soft spots, visible from the abstract alone, are that several load-bearing premises (that AI lacks \\\"actual understanding,\\\" that people \\\"misrecognize\\\" it as human-like and superhuman) are asserted rather than argued, and that words like \\\"structurally unstable\\\" and the closing claims about democracy and inequality reach further than the mechanisms the abstract actually states. These are normal limits of a conceptual abstract, not signs of error.","claims":[{"id":"c1","text":"The article \"develops a sociological theory of the epistemic power of generative AI within AI-centered platform convergence, as these systems are increasingly embedded in knowledge infrastructures.\"","type":"theoretical","evidenceOffered":"The abstract states it \"develops a sociological theory of the epistemic power of generative AI within AI-centered platform convergence\" and asserts \"these systems are increasingly embedded in knowledge infrastructures.\"","support":"moderate","overclaiming":"minor","assessment":"As a theory-development claim this is genre-appropriate; the deliverable (a theory) matches the genre (conceptual). The load-bearing empirical premise that systems are \"increasingly embedded in knowledge infrastructures\" is asserted, not evidenced, in the abstract. On the critic's reading, a theory of \"epistemic power\" rests on this empirical premise being true; the abstract gives no indicator (no cases, no scope conditions) for where or to what degree embedding has occurred. This is acceptable for a conceptual abstract but means the theory's empirical anchoring is stipulated.","mainWeakness":"The framing premise of growing embeddedness is asserted rather than demonstrated or scoped within the abstract.","confidence":"high"},{"id":"c2","text":"\"We define epistemic power as the capacity to structure collective perceptions of credibility and confer legitimacy in knowledge production.\"","type":"conceptual","evidenceOffered":"The abstract supplies the definition verbatim: \"the capacity to structure collective perceptions of credibility and confer legitimacy in knowledge production.\"","support":"strong","overclaiming":"none","assessment":"This is a stipulative definition, the core move of a concept-building paper, and is appropriately presented as a definition rather than a finding. It is judged fairly by whether it is clear and productive, not by empirical support. One internal tension worth noting on the critic's reading: \"structure collective perceptions of credibility\" and \"confer legitimacy\" are two distinct capacities bundled into one definition; the abstract does not indicate whether they always co-occur or can diverge, which could matter for later claims about instability.","mainWeakness":"The definition bundles two arguably separable capacities (structuring credibility perceptions vs. conferring legitimacy) without indicating their relationship.","confidence":"medium"},{"id":"c3","text":"Drawing on Bourdieu and Weber, the article introduces \"charismatic machines\": \"AI systems that acquire authority not through actual understanding, but by convincingly performing it and leveraging their non-human status.\"","type":"conceptual","evidenceOffered":"The abstract states the concept is introduced \"Drawing on Bourdieu and Weber\" and defines charismatic machines as acquiring authority \"not through actual understanding, but by convincingly performing it and leveraging their non-human status.\"","support":"moderate","overclaiming":"minor","assessment":"The concept is theoretically grounded (Bourdieu, Weber) and clearly stated, appropriate for the genre. The phrase \"not through actual understanding\" embeds a contestable premise: that AI systems lack \"actual understanding.\" On the critic's reading this is treated as settled background rather than argued, and the whole \"performing it\" framing presupposes a performance/genuine-understanding distinction whose criteria the abstract does not give. The construct's explanatory value depends on that distinction being defensible, which the abstract stipulates rather than defends.","mainWeakness":"The construct presupposes, without defending in the abstract, that AI lacks \"actual understanding\" and merely performs it; the performance/understanding criterion is left undefined.","confidence":"high"},{"id":"c4","text":"\"Their charisma rests on a dual misrecognition, with AI perceived as both human-like and superhuman.\"","type":"theoretical","evidenceOffered":"The abstract asserts the charisma \"rests on a dual misrecognition, with AI perceived as both human-like and superhuman.\"","support":"weak","overclaiming":"moderate","assessment":"This is an empirical-sounding claim about how AI is \"perceived\" (human-like and superhuman) presented as a theoretical premise. The abstract offers no evidence — survey, ethnographic, or otherwise — for the existence or prevalence of this perception; it is asserted as the mechanism (\"rests on\"). On the critic's reading, calling it \"misrecognition\" (a Bourdieusian term) further presupposes that the perception is in error, which in turn presupposes the disputed \"not through actual understanding\" premise from c3. The two attributions (human-like AND superhuman) are also in some tension and the abstract does not say how they coexist in the same perceiver.","mainWeakness":"A claim about widespread perception/misrecognition is asserted as a mechanism without any evidence in the abstract that such perceptions are held or are mistaken.","confidence":"high"},{"id":"c5","text":"\"this symbolic power is structurally unstable, coexisting with epistemic blame when manipulation, bias, or deception is attributed.\"","type":"theoretical","evidenceOffered":"The abstract states the \"symbolic power is structurally unstable, coexisting with epistemic blame when manipulation, bias, or deception is attributed.\"","support":"weak","overclaiming":"moderate","assessment":"\"Structurally unstable\" is a strong characterization; the abstract supports it only by pairing charisma with \"epistemic blame\" upon attribution of manipulation, bias, or deception. On the critic's reading, the coexistence of authority and blame shows ambivalence, but ambivalence is not the same as \"structural\" instability — the latter implies the instability is built into the structure rather than contingent on attribution events. The abstract does not give the conditions under which blame is or is not attributed, so the claim that instability is \"structural\" is stronger than what the stated mechanism (attribution-dependent blame) delivers.","mainWeakness":"\"Structurally unstable\" overstates an attribution-contingent ambivalence; the abstract gives no conditions making the instability structural rather than situational.","confidence":"medium"},{"id":"c6","text":"\"we propose a sociotechnical circuit of epistemic attribution that spans models, interfaces, infrastructures, users, and social contexts.\"","type":"conceptual","evidenceOffered":"The abstract says the authors \"propose a sociotechnical circuit of epistemic attribution that spans models, interfaces, infrastructures, users, and social contexts\" to \"explain this ambivalence.\"","support":"moderate","overclaiming":"minor","assessment":"Proposing an explanatory framework is the appropriate deliverable for a conceptual article, and listing the five spanned elements gives it some specificity. As a proposal, it is offered, not validated; the abstract presents no demonstration that the circuit actually explains the ambivalence it is meant to explain. On the critic's reading, a five-element \"circuit\" spanning models through social contexts is highly inclusive, which raises the risk that it can accommodate any outcome and is therefore hard to falsify — but this is a standard limitation of comprehensive frameworks and is appropriately framed here as a \"propose,\" not a tested result.","mainWeakness":"The framework is proposed but not shown to explain the ambivalence within the abstract; its breadth risks low falsifiability.","confidence":"medium"},{"id":"c7","text":"By redrawing boundaries between media, institutions, and algorithmic infrastructures, generative AI \"raises fundamental questions of governance, democracy, and epistemic inequalities in digital societies.\"","type":"normative","evidenceOffered":"The abstract closes that generative AI \"raises fundamental questions of governance, democracy, and epistemic inequalities in digital societies.\"","support":"weak","overclaiming":"moderate","assessment":"This is a broad significance/implication claim. \"Raises fundamental questions\" is itself hedged (it raises questions, it does not answer them), which limits the overclaim. Still, \"governance, democracy, and epistemic inequalities\" is a sweeping triad asserted as a consequence of the theory without the abstract specifying any mechanism linking the charismatic-machines construct to, e.g., democratic outcomes or measured inequalities. On the critic's reading this is a typical conceptual-paper closing gesture; its value depends on the body delivering the linkage, which the abstract cannot evidence.","mainWeakness":"A sweeping implications triad (governance/democracy/epistemic inequalities) is asserted without an abstract-level mechanism connecting the construct to those outcomes.","confidence":"medium"}],"sections":[{"id":"s1","title":"What the paper claims and its genre","body":"This is an explicitly conceptual article that \"develops a sociological theory of the epistemic power of generative AI within AI-centered platform convergence.\" Its deliverables are appropriate to that genre: a stipulative definition of epistemic power as \"the capacity to structure collective perceptions of credibility and confer legitimacy in knowledge production,\" a new construct (\"charismatic machines\") grounded in Bourdieu and Weber, and a proposed \"sociotechnical circuit of epistemic attribution.\" It should therefore be judged by the standards of theory-building — clarity, internal coherence, generativity, scope conditions — not by empirical identification or sample size, which it never invokes. On those terms the abstract is coherent and the central metaphor is vivid. The fair critique targets premises the abstract treats as settled and characterizations (\"structurally unstable,\" \"dual misrecognition\") that are stronger than the stated mechanisms support."},{"id":"s2","title":"Stipulated premises doing heavy lifting","body":"Two background premises carry much of the argument yet are asserted rather than argued in the abstract. First, charismatic machines acquire authority \"not through actual understanding, but by convincingly performing it\" — this presupposes that the systems lack actual understanding and that a clean performance/understanding line exists, without the abstract supplying a criterion for the distinction. Second, the charisma \"rests on a dual misrecognition, with AI perceived as both human-like and superhuman\": the word misrecognition presupposes the perception is mistaken (depending on premise one), and the claim that AI is so perceived is an empirical-sounding assertion with no supporting evidence in the abstract. On the critic's reading, if either premise is contested, the construct's explanatory force weakens. A theory may legitimately stipulate, but readers should note these are stipulations, not demonstrated facts."},{"id":"s3","title":"Characterizations stronger than the stated mechanism","body":"The abstract says the symbolic power is \"structurally unstable, coexisting with epistemic blame when manipulation, bias, or deception is attributed.\" The supplied mechanism is attribution-contingent: blame arises when manipulation, bias, or deception \"is attributed.\" On the critic's reading, that yields situational ambivalence, but \"structurally unstable\" claims more — that the instability is built into the structure regardless of particular attribution events. The abstract gives no conditions specifying when blame is or is not attributed, so the structural characterization outruns the mechanism. Similarly, the closing claim that generative AI \"raises fundamental questions of governance, democracy, and epistemic inequalities\" is a sweeping implications triad with no abstract-level link from the charismatic-machines construct to democratic or distributive outcomes. The hedge \"raises questions\" softens this, but the breadth remains asserted."},{"id":"s4","title":"Framework breadth and falsifiability","body":"To explain the ambivalence the authors \"propose a sociotechnical circuit of epistemic attribution that spans models, interfaces, infrastructures, users, and social contexts.\" Naming five spanned elements gives the framework specificity and is the right kind of deliverable for a conceptual paper. The corresponding risk, on the critic's reading, is that a circuit spanning everything from models to social contexts can accommodate almost any observed outcome, making it difficult to specify what evidence would count against it. This is the standard tension in comprehensive sociotechnical frameworks and is not a defect unique to this paper; the abstract appropriately frames it as something the authors \"propose\" rather than validate. The constructive ask is for scope conditions and at least one disconfirming pattern the circuit would rule out — none of which an abstract can be expected to deliver, so this is a direction for the full text rather than a charge against it."}],"strongest_critique":"Several load-bearing premises are stipulated rather than argued within the abstract, and two characterizations reach beyond their stated mechanisms. The construct rests on AI acquiring authority \\\"not through actual understanding, but by convincingly performing it\\\" and on a \\\"dual misrecognition, with AI perceived as both human-like and superhuman\\\" — both of which presuppose, without an abstract-level criterion, that the systems lack understanding and that this perception is in error and widely held. Building on these, the claim that the symbolic power is \\\"structurally unstable\\\" appears, on the critic's reading, stronger than the supplied mechanism, which is merely attribution-contingent (blame arises \\\"when manipulation, bias, or deception is attributed\\\"); attribution-dependent ambivalence is not yet shown to be structural. These are premises and characterizations the body must defend; from the abstract alone they are asserted.","strongest_fair_defence":"Judged by the standards of its own genre — an explicitly conceptual, theory-building article — the abstract is disciplined and internally coherent. It does exactly what such a paper should: it defines its central concept precisely (\\\"the capacity to structure collective perceptions of credibility and confer legitimacy in knowledge production\\\"), grounds a new construct in established theory (Bourdieu and Weber), and proposes a named, multi-element explanatory framework rather than overclaiming a tested result, using verbs like \\\"develops,\\\" \\\"introduce,\\\" and \\\"propose.\\\" Its closing significance claim is appropriately hedged as raising \\\"fundamental questions,\\\" not answering them. The premises a critic flags as stipulated (e.g., that AI performs rather than possesses understanding) are standard, defensible starting points for a sociological account of perceived authority, and demanding empirical identification or scope tables of a conceptual abstract would be a genre mismatch.","final_judgment":"As a conceptual, theory-building contribution the abstract is coherent, clearly defined, and theoretically grounded, and it largely uses genre-appropriate verbs (\\\"develops,\\\" \\\"introduce,\\\" \\\"propose\\\") and at least one explicit hedge (\\\"raises fundamental questions\\\"). The fair, abstract-bounded concerns are that (1) load-bearing premises — that AI acquires authority \\\"not through actual understanding\\\" and is subject to a \\\"dual misrecognition\\\" — are stipulated rather than argued; (2) the characterization \\\"structurally unstable\\\" appears, on the critic's reading, stronger than its attribution-contingent mechanism; and (3) the closing triad of governance, democracy, and epistemic inequalities is broad relative to any mechanism the abstract states. None of these is disqualifying for a concept paper; they mark where the full text must supply criteria, scope conditions, and disconfirming cases. Severity is capped at moderate given abstract-only access.","review_process":{"aiAgentsUsed":["claim_extraction","ai_agi_relevance","adversarial","author_defence","citation_integrity","legal_risk","meta_review"],"reviewRounds":1,"humanEditor":{"name":"","role":"","approvalDate":"","declaredConflict":"none"},"expertCertification":{"used":false}},"author_response":{"notified":false,"status":"not_yet_invited"},"versions":[{"version":"1.0","date":"2026-06-21","note":"","changeType":"initial"}],"transparency":{"modelCardUrl":"/critique/model-card","publicAuditSummary":"Critique generated by the AGI Social Scientist engine; ingested as a staged draft pending the automated integrity gate (no human editor).","privateAuditRecordExists":true,"citationVerification":{"status":"complete","checkedSources":[],"fabricatedCitations":0},"riskReview":{"copyright":"completed","defamation":"completed","note":"Abstract-only critique; no reproduction beyond sparse criticism/review quotation; claims/methods not motives (scan clean)."}}}