{"id":"adccf565-db91-4fce-a27f-bdff97eff516","arxiv_id":"2507.01483","paper_version":1,"verdict":"REJECT","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"high","formal_verification":"none","parameter_count":0,"one_line_summary":"An Austrian-school essay argues AI systems cannot perform core economic coordination tasks and introduces 'epistemic scarcity' as a new primitive.","lead":"This paper argues that artificial intelligence cannot replace human economic judgment because only people can interpret ends, discover means, and communicate subjective value. It proposes 'epistemic scarcity', the idea that too much information makes reliable knowledge harder to find, as the central economic problem of the AI age.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The paper's categorical claim that AI cannot perform economic coordination is unsupported: §5's Scarcity Index uses |Ω| for a state space the paper itself declares non-enumerable, and §4.2's 'select vs choose' dichotomy is a stipulative definition, leaving the central claim unfalsifiable.","rationale":"The paper is best read as an Austrian-school philosophical essay; its Mises/Hayek citations are genuine, and it is candid about its a priori method. I agree with the reader's REJECT verdict, but I locate the load-bearing problem one step deeper. The reader's weakest assumption targets the undefined Scarcity Index, which is a real and concrete defect; however, that defect mainly undermines the 'new primitive' side-claim. The central claim about AI is more severely compromised because it inherits the same undefined formal apparatus and then adds a stipulative definitional split (§4.2) between selection and choice, which makes the categorical conclusion immune to disproof. The paper does not provide a non-computability argument in any standard sense (e.g., a proof that 'interpretation of ends' is not Turing-computable); it asserts an ontological boundary. Its own case study of AI-driven market moves (§8.2) shows AI can influence prices, so the categorical claim survives only by excluding such influence as inauthentic communication, a criterion no measurement could contradict. Because the claims are universal, categorical, and unfalsifiable, and because the formal model is ill-defined, the REJECT verdict stands unchanged. I would note that the paper has independent value as an opinion/policy statement within its tradition, but it does not meet the evidentiary bar for its strongest empirical-sounding assertions.","tokens_in":32232,"tokens_out":7305,"duration_ms":85385,"concrete_test":"Run a finite-state instantiation of the §5.2 model: fix Ω as an explicit set of N possible world-states (e.g., 10 binary market-relevant features) and compute S_i = 1 − |Dom(P_i)| / |Ω| for a simulated agent. If the index is computable and behaves as claimed for finite Ω, then the paper's 'indefinable Ω' is a stylistic claim, not a formal one, and the model does not establish a categorical boundary; if the index cannot be computed even for this finite case, the model is vacuous at any scale. Either outcome removes the formal support for the paper's categorical AI-incapability thesis.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that AI systems are categorically incapable of interpreting ends, discovering means, and communicating subjective value is defended with a formal model whose key quantities are undefined. Section 5.1 asserts that Ω is 'indefinable' and 'non-enumerable,' yet the Scarcity Index in §5.2 is defined as 1 − |Dom(P_i)| / |Ω|, which requires cardinalities of a set that has no cardinal number; Section 13.3 concedes the construct 'resists reduction to static data points' and cannot be operationalized. The model therefore cannot draw the boundary between ordinary uncertainty and epistemic scarcity on which the human/machine distinction is mounted. Furthermore, the claim is unfalsifiable as stated: §4.2 says 'machines can select; they cannot choose' without any operational test or impossibility proof, making the conclusion analytic rather than demonstrated, and the paper's own case material (§8.2, AI-generated headlines moving markets) shows AI causally affecting price signals, which the paper dismisses as simulacra only by appeal to a non-empirical criterion of authenticity.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces \"epistemic scarcity\" as a new economic primitive, formalizes it with a Scarcity Index, and uses it to argue that AI systems are categorically incapable of interpreting ends, discovering means, and communicating subjective value through market signals. It also critiques FAT-style AI ethics from an Austrian praxeological standpoint, proposes epistemic property rights and related governance mechanisms, and concludes with a civilizational claim about the defense of truth. The argument relies heavily on Misesian apriorism, Hayekian coordination, and Hoppe's argumentation ethics.","tokens_in":32481,"tokens_out":5450,"duration_ms":66723,"significance":"If the central claim were established, the paper would mark a sharp boundary on the scope of AI in economic coordination and would challenge much of current algorithmic governance and AI ethics research. The paper does synthesize a wide literature and raises substantive questions about algorithmic opacity, institutional trust, and the epistemology of markets. However, the formal model is not well-formed, the central distinction is stipulated rather than demonstrated, and the paper itself concedes that its core construct resists operationalization. Those problems are load-bearing because the categorical AI conclusion rests on the formalism and on the select/choose dichotomy. The manuscript therefore does not currently deliver the significance it promises, although the questions it poses are worth further study.","major_comments":[{"comment":"The formal model is internally inconsistent. Section 5.1 declares Ω 'indefinable' and 'non-enumerable,' yet Section 5.2 defines the Scarcity Index as S_i = 1 − |Dom(P_i)|/|Ω|, which requires cardinalities for Ω and Dom(P_i). If Ω is indefinable, its cardinality is not defined and the index is not a well-formed quantity. The condition 'Im(f) ⊂ ∅' in Section 5.1 is also not a substantive condition: if f is undefined, Im(f) is undefined, and if f is defined on a nonempty Ω, Im(f) is nonempty and cannot be a subset of the empty set. Section 13.3 concedes that epistemic scarcity 'resists reduction to static data points' and cannot currently be operationalized; that concession does not repair the formalism, because the problem is that the formal quantities are undefined, not merely difficult to measure. As a consequence, the Scarcity Index cannot support the asserted distinction between ordinary uncertainty and epistemic scarcity on which the paper's human/machine conclusion depends.","section":"§5.1–5.2"},{"comment":"The central claim that machines can 'select' but not 'choose,' and cannot 'originate conceptual categories,' is stipulated rather than demonstrated. No operational test, impossibility proof, or account of what evidence would count against the claim is provided. Because the conclusion is embedded in the definitions of 'choose,' 'value,' and 'end,' the case studies cannot refute it; they can only be interpreted as simulacra. An unfalsifiable thesis of this kind cannot carry the abstract's categorical assertion that AI is 'categorically incapable' of performing economic coordination. A weaker claim about current AI systems might be defensible, but the paper does not provide the argument needed to support the stronger claim.","section":"§4.2 and §5.4.1"},{"comment":"The paper's own evidence works against its categorical conclusion. Section 6.2 cites financial-market fluctuations triggered by AI-generated false headlines, and Section 8.2 treats AI hallucinations as economically consequential. These examples show algorithmic outputs affecting price signals and market beliefs. To reconcile this with the claim that AI cannot 'communicate subjective value through market signals,' the paper invokes a distinction between authentic signals and simulacra, but no criterion for authenticity is defined independently of the conclusion. This is a load-bearing gap: if AI-generated content moves prices and coordinates expectations, then the claim that AI is categorically outside economic coordination requires a stronger argument than the stipulated select/choose distinction.","section":"§6.2 and §8.2"},{"comment":"The use of Hoppe's argumentation ethics to declare all denials of the action axiom 'self-refuting' is a dialectical strategy that insulates the framework from empirical evidence. Even if the performative-contradiction argument were accepted for the action axiom, the paper does not show how it licenses the additional premises that ends are non-computable (Section 4.3) and that machines cannot originate norms (Section 12.1). Those premises are doing the work in the AI conclusion, and they are asserted rather than proved. The reader is therefore asked to accept the essay's central claim on the basis of definitions, not on the basis of the formal model or case studies.","section":"§3.1 and §12.1"}],"minor_comments":[{"comment":"There are numerous typographical errors, including the missing space in 'challengingprevailingassumptions' in the abstract and the misspelling of 'façade' in Section 8.3; the manuscript needs careful proofreading.","section":"§1 and throughout"},{"comment":"The expression 'Es = ∂K/∂C' is presented without derivation and is not connected to the Scarcity Index S_i; the paper should either integrate these formal proposals or remove the unexplained derivative notation.","section":"§1.3 and §5"},{"comment":"The phrase '∃ epistemic scarcity' is malformed notation and should be replaced with prose.","section":"§5.1"},{"comment":"The limitations section is honest, but it directly contradicts the formal claims made in Section 5; the paper should acknowledge that contradiction and explain how the formal model can be interpreted if epistemic scarcity cannot be operationalized.","section":"§13.3"},{"comment":"Reference [2] is an unpublished MSc dissertation cited repeatedly as empirical authority for load-bearing claims about Dark Triad traits and manipulation; the paper should rely on independent, peer-reviewed evidence for such claims.","section":"Reference [2]"}],"recommendation":"reject","confidential_remarks":"The paper is a programmatic essay in which the formal economics contribution is not well-formed and the central conclusion is unfalsifiable as stated. The repeated self-citation of Reference [2] for empirical claims and the lack of engagement with the possibility that AI could instantiate the relevant economic functions make the manuscript unsuitable for publication in its current form. The topic is interesting, but the central claim would need to be substantially weakened and the formal model either removed or properly defined before the paper could be considered."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"What you should know: this is a clearly written Austrian-school essay, not a research result. It repackages Hayek, Lachmann, Shackle, and the post-truth literature under the label 'epistemic scarcity' and argues that AI cannot do the interpretive work of economic coordination. The abstract promises a categorical result; the paper delivers an assertion derived from its own premises.\n\nWhat it does well: the four-part typology of opacity—stochastic, semantic, strategic, recursive—is a handy organizing device. The case studies (COVID expert consensus, LLM hallucination, OgAS/Soviet cybernetics) are readable and do real illustrative work. The paper also engages the Austrian canon seriously, and Section 13.3 is admirably honest that the central concept resists operationalization. If you want a policy-adjacent restatement of Austrian objections to algorithmic governance, this is a competent specimen.\n\nWhere it falls apart: the formal model in Section 5. The paper declares the state space Ω 'indefinable' and 'non-enumerable,' then defines the Scarcity Index as S_i = 1 − |Dom(P_i)|/|Ω|. Cardinalities of an indefinable, non-enumerable set are not defined, so the index is a symbol, not a quantity. Section 13.3 concedes exactly this. The 'select vs choose' distinction in Section 4.2 is stipulative, which makes the central claim unfalsifiable. And Section 8.2 shows AI-generated headlines moving markets, then dismisses those as simulacra only by appealing to a non-empirical criterion of authenticity. That is not a demonstrated result; it is a worldview.\n\nThe citation pattern is another soft spot. Some references look plausible but do not resolve when checked against their listed venues, and the paper leans on the author's own MSc dissertation for empirical support. Self-citation is not by itself a flaw, but here it is doing a lot of work alongside references I could not verify.\n\nWho should read it: people who want a clearly written Austrian-theory brief on AI and epistemic governance. It is a policy statement, not a contribution to economic theory or AI research. I would not cite it in my own work, and I would not send it to peer review as a serious research submission. A desk reject is appropriate, unless an editor specifically wants a methodological referee to confirm that the formal model is not salvageable.","headline":"A fluent Austrian polemic that names 'epistemic scarcity' but never earns its categorical AI claim; the formal model trips on the undefined state space, and several citations don't check out.","tokens_in":33007,"tokens_out":2216,"would_cite":false,"duration_ms":30075,"reading_group":"maybe","serious_thinker":"no","would_accept_peer_review":false},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper argues that AI cannot perform the central functions of economic coordination — interpreting ends, discovering means, signalling subjective value — and names 'epistemic scarcity' as the new binding market constraint.","keywords":["epistemic scarcity","Austrian economics","praxeology","AI and economic coordination","entrepreneurial foresight","algorithmic governance","subjective value","post-truth markets"],"falsifier":"A controlled market experiment would settle the categorical claim: an autonomous system given no human-assigned objective beyond a general resource constraint would have to originate an end, discover a profit opportunity that no human participant anticipates, and realise a profit by trading against real counterparties — one success falsifies the thesis that AI cannot interpret ends and discover means. A second, softer test is statistical: if epistemic scarcity is real, proxies such as volatility in expert consensus, reputational entropy, or belief-updating lag should measurably predict coordination failure (for example, wider price dispersion) in low-verifiability markets, and the paper's Scarcity Index should be computable for real agents, which Section 13.3 concedes it currently cannot be.","tokens_in":31970,"feed_emoji":"🤖","tokens_out":13209,"duration_ms":126662,"temperature":0.7,"pith_summary":"This paper tries to establish that AI systems are categorically incapable of performing the central functions of economic coordination: interpreting human ends, discovering the means to satisfy them, and communicating subjective value through prices. Its diagnosis is that contemporary markets suffer not from scarce information but from 'epistemic scarcity' — the condition in which the marginal cost of obtaining verifiable, decision-relevant knowledge dominates every other cost, because truth is not hidden but drowned in engineered noise. Drawing on the a priori theory of action at the root of Austrian economics, the paper argues that entrepreneurial foresight is an imaginative act of creating possible futures rather than an extrapolation from data, so machines can select but cannot choose, can optimise but cannot value. It formalises the condition with a scarcity index and a four-part typology of opacity, then draws the political conclusion that algorithmic governance, like Soviet cybernetics, mistakes calculation for coordination. If the argument holds, the debate over AI is not a technical dispute but a choice between computational social control and reasoned human autonomy.","feed_headline":"AI cannot do the core work of markets, an Austrian analysis argues","feed_subtitle":"Machines can mimic choice but never originate ends, discover means, or signal value through prices.","key_machinery":"Two devices carry the argument. The first is the action axiom: the Misesian claim, sharpened by Rothbard and Hoppe, that all human conduct is purposive action and that this is a synthetic a priori truth no data can displace — which is what makes the paper's capability claim categorical rather than empirical. The second is the epistemic scarcity formalism: a gradient $E_s = \\partial K / \\partial C$ measuring the cognitive or institutional cost of acquiring accessible knowledge, with $\\partial^2 K / \\partial C^2 < 0$, and a Scarcity Index $S_i = 1 - |\\mathrm{Dom}(P_i)| / |\\Omega|$, where $\\Omega$ is the state space of relevant world conditions, held to be indefinable and non-enumerable under imposed opacity. The paper's typology of opacity — stochastic (the world outpaces cognition), semantic (signals lose their referents), strategic (noise is injected adversarially), and recursive (the trustworthiness of frames is itself unknowable) — does the load-bearing work of separating epistemic scarcity from ordinary Knightian uncertainty: under epistemic scarcity the agent cannot construct the state space over which probabilities would range, so Bayesian updating has nothing to update.","core_discovery":"The paper's central claim is that the core functions of economic coordination — the interpretation of ends, the discovery of means, and the communication of subjective value through market signals — are irreducibly human acts, and that algorithmic systems, however complex, cannot originate normativity, interpret institutional change, or bear responsibility. It frames decision-making not as optimisation under constraint but as purposive action under uncertainty, and it treats the price system as an epistemic artefact whose coordinating power depends on the semantic fidelity of its signals; when signal integrity collapses under strategic obfuscation, spontaneous order degenerates into a simulacrum of order. To name this new condition it introduces epistemic scarcity as a new economic primitive, formalised as the gradient $E_s = \\partial K / \\partial C$ with diminishing returns to epistemic investment, and a Scarcity Index $S_i = 1 - |\\mathrm{Dom}(P_i)| / |\\Omega|$ that tends to unity as the agent's domain of coherent belief collapses. The paper's closing claim is civilisational: if AI cannot perform the interpretive labour of markets, then attempts to automate economic coordination — from AI ethics frameworks to algorithmic planning — replicate the error of the Soviet cybernetics project, and the Austrian tradition of action, subjectivity, and spontaneous order is the only coherent alternative to computational control.","pith_inferences":["A weaker, testable version of the categorical claim follows from the paper's own terms: AI should perform well in economic domains with stable, enumerable state spaces (logistics, pricing under risk, games) and fail where ends must be interpreted, so the boundary between computable and non-computable economic function could be mapped empirically domain by domain even though the paper states its th","If epistemic scarcity is a genuine economic primitive, then the proxy measures the paper lists as embryonic — volatility in expert consensus, reputational entropy, lag times in belief updating — should predict coordination failure, such as wider price dispersion, in low-verifiability markets like finance, media, and health care; this is an empirical extension the author leaves to future work.","The argument implies a strategic asymmetry the paper does not draw out: in an economy of manufactured opacity, the scarce resource is not information but trustworthy interpretation, so investment in verification infrastructure — cryptographic provenance, adversarial reputation markets, staked truth-bonding — becomes a profit opportunity rather than a regulatory cost.","The framework predicts that verification will rival production as the site of economic activity: as AI lowers the cost of plausible falsehood, human intermediaries who stake their reputations on claims should earn scarcity rents, a structural shift the paper gestures at with its 'epistemic entrepreneurs' but does not quantify."],"forward_implications":["No AI system can function as an economic actor, entrepreneur, or moral agent; proposals for AI-run markets, algorithmic monetary policy, and computational social credit inherit the failure of Soviet cybernetics because they mistake calculability for coordination.","Ethical AI frameworks built on fairness, accountability, and transparency metrics embody constructivist rationalism: they impose paternalistic constraints and violate voluntary exchange, so attempts to encode morality in code misunderstand both ethics and economics.","Information abundance becomes a source of market failure: when verification costs rise asymptotically relative to the cost of producing noise, price signals decouple from underlying realities and coordination fails from semiotic excess, not ignorance.","A new class of epistemic entrepreneurs emerges — agents who profit by verifying, authenticating, and curating claims — making reputation markets and trusted signalling essential infrastructure of exchange.","Because AI cannot bear responsibility, liability for algorithmic harm falls to the human actors who design and deploy the systems, which the paper develops into proposals for epistemic property rights and adversarial verification rather than centralised speech regulation."],"supporting_citations":[{"why":"Supplies the Misesian action axiom from which the paper derives the categorical claim that purposive ends-interpretation is irreducibly human.","marker":"[17]"},{"why":"Provides the Hayekian price-signal mechanism that the paper argues collapses when signal integrity fails and that AI allegedly cannot replace.","marker":"[16]"},{"why":"Grounds the distinction between historical understanding and causal explanation that the paper uses to indict algorithmic inference as epistemically blind.","marker":"[53]"},{"why":"Supplies argumentation ethics, used to show that empirical and constructivist challenges to a priori action categories are performatively self-refuting.","marker":"[20]"},{"why":"Contributes the kaleidic-structures argument that the economic environment has no stable sample space, undermining probabilistic forecasting by AI.","marker":"[25]"},{"why":"Supports the claim that entrepreneurial decision is the imaginative creation of possible futures, not selection among known outcomes.","marker":"[26]"},{"why":"The author's own empirical study of Dark Triad traits, used as evidence that epistemic asymmetries are strategically exploited in ambiguous environments.","marker":"[2]"},{"why":"Documents that falsehoods spread faster than truths, supplying the empirical anchor for the paper's mechanism of epistemic erosion.","marker":"[34]"},{"why":"The computational-planning proposal the paper rebuts, representing the position that algorithmic coordination can replace markets.","marker":"[52]"}],"fun_headline_variants":["AI can't signal value: markets need human interpretation","Epistemic scarcity: why AI can't replace market coordination","Austrian view: AI can mimic choice, never originate ends","AI's epistemic limits: no price system without human action","Markets need humans: AI can't discover means or ends"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise, asserted from the action axiom rather than demonstrated, is that interpreting ends, discovering means, and valuing are categorically human acts that no algorithm could ever instantiate — and the paper's own formal index presupposes well-defined state and belief sets ($\\Omega$ and $\\mathrm{Dom}(P_i)$) that its central argument declares indefinable and non-operationalisable.","fun_headline_variants_meta":{"raw":{"variants":["AI can't signal value: markets need human interpretation","Epistemic scarcity: why AI can't replace market coordination","Austrian view: AI can mimic choice, never originate ends","AI's epistemic limits: no price system without human action","Markets need humans: AI can't discover means or ends"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000186,"raw_usage":{"total_tokens":1377,"prompt_tokens":1047,"completion_tokens":330,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":663,"completion_tokens_details":{"reasoning_tokens":247}},"tokens_in":663,"tokens_out":330,"duration_ms":4246,"temperature":1.0,"reasoning_tokens":247,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T20:50:04.369280+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A controlled market experiment would settle the categorical claim: an autonomous system given no human-assigned objective beyond a general resource constraint would have to originate an end, discover a profit opportunity that no human participant anticipates, and realise a profit by trading against real counterparties — one success falsifies the thesis that AI cannot interpret ends and discover means. A second, softer test is statistical: if epistemic scarcity is real, proxies such as volatility in expert consensus, reputational entropy, or belief-updating lag should measurably predict coordination failure (for example, wider price dispersion) in low-verifiability markets, and the paper's Scarcity Index should be computable for real agents, which Section 13.3 concedes it currently cannot be.","supporting_citations":[{"cited_title":"Human Action: A Treatise on Economics","cited_arxiv_id":null,"evidence_quote":"Supplies the Misesian action axiom from which the paper derives the categorical claim that purposive ends-interpretation is irreducibly human."},{"cited_title":"Theory and History: An Interpretation of Social and Economic Evolution","cited_arxiv_id":null,"evidence_quote":"Grounds the distinction between historical understanding and causal explanation that the paper uses to indict algorithmic inference as epistemically blind."},{"cited_title":"A Theory of Socialism and Capitalism","cited_arxiv_id":null,"evidence_quote":"Supplies argumentation ethics, used to show that empirical and constructivist challenges to a priori action categories are performatively self-refuting."},{"cited_title":"Lachmann","cited_arxiv_id":null,"evidence_quote":"Contributes the kaleidic-structures argument that the economic environment has no stable sample space, undermining probabilistic forecasting by AI."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supports the claim that entrepreneurial decision is the imaginative creation of possible futures, not selection among known outcomes."},{"cited_title":"Dark Triad Traits, Epistemic Exploitation, and Strategic Manipula- tion","cited_arxiv_id":null,"evidence_quote":"The author's own empirical study of Dark Triad traits, used as evidence that epistemic asymmetries are strategically exploited in ambiguous environments."},{"cited_title":"Towards a New Socialism","cited_arxiv_id":null,"evidence_quote":"The computational-planning proposal the paper rebuts, representing the position that algorithmic coordination can replace markets."}],"review_version":1}