{"$schema":"https://policywindow.org/critique/api/schema","critique_id":"CRIT-000045","slug":"cultural-tendencies-generative-ai","url":"https://policywindow.org/critique/c/cultural-tendencies-generative-ai","doi":null,"status":"published","critique_type":"editorially_approved_ai_native_critique","publication_date":"2026-07-05","current_version":"1.1","target_paper":{"title":"Cultural tendencies in generative AI","authors":["Jackson G. Lu","Lesley Luyang Song","Lu Doris Zhang"],"journal":"Nature Human Behaviour","doi":"10.1038/s41562-025-02242-1","url":"https://doi.org/10.1038/s41562-025-02242-1","publicationDate":"2025","paperType":"empirical","accessBasis":"licensed_access","fullTextUsed":true,"fictional":false,"doi_url":"https://doi.org/10.1038/s41562-025-02242-1"},"source_journal":{"tier":"A","rankingSources":["resolved from the monitored-venue determination"],"rankingNote":"A-tier per the monitored-venue determination; Nature Human Behaviour is elite interdisciplinary (Nature family), WoS JCR Q1. Critiqued from the publisher's version of record under the operator's licensed subscription access (not a paywall bypass)."},"selection_provenance":{"id":"cultural-tendencies-generative-ai","venue":"Nature Human Behaviour","inMonitoredSet":true,"determinedTier":"A","recordedTier":"A","effectiveTier":"A","kind":"monitored","disclosed":true,"offListPeerReviewed":false},"selection":{"aiAgiCentralityScore":4,"societalRelevanceScore":5,"aiAgiCategories":["human_AI_interaction"],"selectionReason":"Autonomous production cycle (licensed inbox, A-tier priority); full-text critique via two-stage produce+sharpen + 3-lens convergence gate (3 survives, 0 weakened).","domain":"psychology"},"scores":{"aiAgiContribution":4,"evidentiarySupport":4,"methodologicalRisk":2,"overclaiming":3,"reproducibilityOrAuditability":3,"societalImpactRelevance":5,"severity":"moderate","confidence":"high"},"severity_cap_for_access_basis":"high","plain_language_summary":"This study prompts GPT-4 and ERNIE in Chinese versus English on established cultural-psychology scales (collectivism, attribution bias, reasoning style, expectation of change), finding that Chinese-language prompts yield more interdependent social-orientation scores and more holistic cognitive-style scores. It then shows GPT recommends culturally congruent advertising slogans and that cultural role-prompts can shift outputs. The design is descriptive and correlational, comparing 100 API iterations per language condition on single frozen model snapshots. The central critique is that the paper labels GPT's differential ad-slogan recommendations as 'real-world impact' when no human consumers, attitudes, or behavioural outcomes were measured — a model choosing different outputs by language is a model-behaviour finding, not demonstrated impact on people. A secondary concern is that the proposed causal mechanism ('cultural tendencies embedded in large-scale textual data') is underdetermined: prompt language co-varies with training-data composition, tokeniser behaviour, and RLHF tuning, none of which the design can separate.","claims":[{"id":"CLAIM-001","text":"The paper claims to 'demonstrate the real-world impact' of cultural tendencies based on GPT's differential ad-slogan recommendations, but the study measures only which slogan GPT selects — no human consumers, attitudes, or behavioural outcomes are measured.","type":"causal","evidenceOffered":"we demonstrate the real-world impact of these cultural tendencies","support":"weak","overclaiming":"major","assessment":"The advertising study (Exploratory Analyses I) records GPT's slogan choice across 100 iterations per language for three product pairs. No human participants view, evaluate, or act on these recommendations. Demonstrating that a model's output differs by language is a model-behaviour finding. Calling it 'real-world impact' elides the entire causal chain from model recommendation to human behaviour. The paper frames this as 'evidence for the real-world impact' in the Results and uses 'we demonstrate' in the abstract, which is stronger than the data supports.","mainWeakness":"Labelling a model-output difference as 'real-world impact' when no human outcome is measured is a direct scope over-claim.","confidence":"high","anticipatedRebuttal":"The authors could argue the advertising task is a face-valid proxy for real-world consequences — marketers do use LLMs to draft copy, so a systematic language-conditioned slogan bias plausibly propagates into materials people eventually see.","response":"Even granting that LLMs are used in marketing, the study measures neither downstream deployment nor any human response; a proxy's real-world impact must be demonstrated, not assumed, and 'we demonstrate the real-world impact' asserts the very conclusion the proxy was meant to support. The gap is not that impact is implausible but that it is unmeasured.","falsificationCondition":"A study linking the language-conditioned outputs to an actual human outcome — e.g. a randomised test in which people exposed to the differing slogans show measurably different attitudes or choices — would convert this from a model-behaviour finding into demonstrated real-world impact and retire the objection."},{"id":"CLAIM-002","text":"The paper proposes that observed Chinese-vs-English output differences 'probably originated from real-world cultural tendencies embedded in large-scale textual data' without ruling out alternative explanations such as tokeniser effects, RLHF cultural steering, or test-item memorisation.","type":"causal","evidenceOffered":"their cultural tendencies probably originated from real-world cultural tendencies embedded in large-scale textual data, on which generative AI models are trained","support":"moderate","overclaiming":"moderate","assessment":"The design manipulates only prompt language. Training-data composition, tokeniser segmentation, RLHF preference tuning, and potential memorisation of well-known psychometric items all co-vary with language. No ablation, probing study, or training-data analysis is conducted. The cross-model replication (GPT and ERNIE) partially addresses some confounds but both models face the same set of confounds. The hedge 'probably' mitigates but does not eliminate the concern.","mainWeakness":"Proposing a single causal pathway without acknowledging competing explanations overstates what a language-manipulation design can identify.","confidence":"high","anticipatedRebuttal":"The authors hedge with 'probably' and replicate the pattern across two independently developed models (GPT and ERNIE) from different countries, which they could argue makes a shared training-data origin the most parsimonious explanation for the convergent, culturally patterned differences.","response":"Cross-model replication rules out a single vendor's idiosyncrasy but not the confounds common to both: both models tokenise Chinese and English differently, both undergo RLHF, and both may have encountered the same well-known psychometric items — so convergence is equally consistent with the training-data account and with its alternatives. 'Probably' correctly signals uncertainty, but the Discussion still advances the single mechanism without the ablation or probing evidence that would license it.","falsificationCondition":"A design that holds language fixed while varying the suspected confound — tokeniser-controlled prompts, a training-data audit, or activation-probing that localises the effect to learned cultural content rather than segmentation or RLHF — would isolate the mechanism and directly support or refute the training-data claim."},{"id":"CLAIM-003","text":"The paper's title generalises from two specific model snapshots (gpt-4-1106-preview and ERNIE-3.5-8K-0205) to 'generative AI' as a category, though model behaviour can shift substantially across versions.","type":"descriptive","evidenceOffered":"Cultural tendencies in generative AI","support":"moderate","overclaiming":"moderate","assessment":"The title and abstract frame findings as properties of 'generative AI' without version qualification. Model versions are updated frequently; the specific API endpoints used are already deprecated. The Limitations section does acknowledge the need to test other models (Claude, DeepSeek, Gemini) but does not flag the fragility of version-specific findings or that even the next GPT point-release could eliminate all observed effects.","mainWeakness":"Generalising from two dated checkpoints to the category 'generative AI' overstates external validity.","confidence":"moderate","anticipatedRebuttal":"The authors could note that the Limitations section already calls for testing other models, and that a title naming specific checkpoints would be unwieldy and against a field convention that routinely generalises from tested systems to the category.","response":"Acknowledging the limitation in the body does not license the unqualified categorical claim in the title and abstract — the parts most readers and the press actually see — especially when the specific endpoints tested are already deprecated, so the strongest form of the claim may not replicate on any currently deployed model.","falsificationCondition":"Reproduction of the same language-conditioned cultural pattern across several current-generation models (e.g. GPT-4o/next, Claude, Gemini, DeepSeek) would justify the categorical 'generative AI' framing; failure on newer checkpoints would confirm the finding is version-specific."}],"sections":[],"strongest_critique":"The paper claims that the advertising study 'demonstrate[s] the real-world impact' of cultural tendencies, but the study records only which slogan GPT selects across 100 API iterations, with no human consumers, attitudes, or behavioural outcomes measured. A model choosing different outputs in different languages is a model-behaviour finding, not evidence of impact on people. The Discussion compounds this by extrapolating to societal-level cultural divergence ('English-speaking AI users may gradually become more independent and analytic') without any supporting evidence of human behavioural change.","strongest_fair_defence":"The authors are transparent that their claims are descriptive ('our findings are descriptive rather than prescriptive'), explicitly state they 'do not suggest that generative AI models possess cultural tendencies like humans do,' hedge the training-data mechanism with 'probably,' employ multiple converging measures across diverse formats (Likert scales, vignette tasks, imagery tasks, text analysis), replicate across two independently developed models from different countries (GPT and ERNIE), conduct four robustness checks including temperature variation and prompt-format sensitivity, share data and code on OSF, use appropriate statistical tests (t-tests, Poisson regression, beta regression) with power analysis, and take concrete steps to mitigate test contamination by modifying item content and obtaining unpublished materials directly from original researchers. The core descriptive finding — that LLM outputs systematically differ by prompt language in culturally patterned ways — is well-documented even if the causal mechanism remains open.","final_judgment":"A well-executed descriptive study with converging evidence across measures and models, weakened primarily by labelling a model-output difference as 'real-world impact' without any human-outcome data, and secondarily by proposing a single causal mechanism that its language-manipulation design cannot isolate from confounds. These are bounded overclaims on an otherwise rigorous and transparent study that honestly discloses its sampling and model limitations.","review_process":{"aiAgentsUsed":["AGISS critique engine (autonomous production cycle)"],"reviewRounds":1,"humanEditor":{"name":"","role":"","approvalDate":"","declaredConflict":"none"},"expertCertification":{"used":false}},"author_response":{"notified":false,"status":"not_yet_invited"},"versions":[{"version":"1.0","date":"2026-07-05","note":"","changeType":"initial"},{"version":"1.1","date":"2026-07-05","note":"Reformatted into the dialectical Matters-Arising Comment genre (G146): added a structured abstract, a bottom-line verdict, and for each objection an anticipated author reply, our response, and an explicit falsification condition ('what would change our mind'). The verdict, claims, severity, and verbatim spans are unchanged; this revision adds the paper's strongest defences and the Comment's own defeaters, and changes presentation — it does not add adverse findings.","changeType":"major_interpretive"}],"transparency":{"modelCardUrl":"/critique/model-card","publicAuditSummary":"Critique produced by the autonomous production cycle (two-stage produce+sharpen + 3-lens convergence gate, 3 survives / 0 weakened) and auto-published under the operator's auto-publish + post-audit model; the Mon/Thu audit is the post-hoc gate.","privateAuditRecordExists":true,"citationVerification":{"status":"complete","checkedSources":[],"fabricatedCitations":0},"riskReview":{"copyright":"completed","defamation":"completed","note":"Nature Human Behaviour (Springer Nature) quoted sparingly under criticism/review; critique targets claims, methods and inference only. Full text accessed under the operator's licensed subscription (never a paywall bypass)."}}}