{ "format_version": 3, "claim_formal": { "subject": "AI tool usage by humans", "property": "associated with diminished critical thinking and problem-solving abilities", "operator": ">=", "operator_note": "The claim as stated is a universal causal assertion. We interpret it as: at least 3 independent, peer-reviewed or authoritative sources report that AI tool usage is associated with reduced critical thinking or problem-solving abilities. This is a consensus-of-evidence interpretation \u2014 the claim is PROVED if the weight of independent evidence supports the association, even though individual studies show correlation rather than proven causation. Important nuance: the evidence shows this effect is moderated by usage patterns, task stakes, and user confidence \u2014 heavy/uncritical use drives the decline, not all AI usage universally.", "threshold": 3, "proof_direction": "affirm" }, "claim_natural": "Using AI tools makes humans worse at critical thinking and original problem-solving.", "evidence": { "B1": { "type": "empirical", "label": "Gerlich (2025): Negative correlation (r=-0.68) between AI usage and critical thinking scores in 666 participants", "sub_claim": null, "source": { "name": "PsyPost report on Gerlich (2025), Societies 15(1):6", "url": "https://www.psypost.org/ai-tools-may-weaken-critical-thinking-skills-by-encouraging-cognitive-offloading-study-suggests/", "quote": "Participants who reported heavy reliance on AI tools performed worse on critical thinking assessments compared to those who used these tools less frequently." }, "verification": { "status": "verified", "method": "full_quote", "coverage_pct": null, "fetch_mode": "live", "credibility": { "domain": "psypost.org", "source_type": "unknown", "tier": 2, "flags": [], "note": "Unclassified domain \u2014 verify source authority manually" } }, "extraction": { "value": "verified", "value_in_quote": true, "quote_snippet": "Participants who reported heavy reliance on AI tools performed worse on critical" } }, "B2": { "type": "empirical", "label": "Lee et al. (2025, CHI): Higher confidence in GenAI associated with less critical thinking in 319 knowledge workers", "sub_claim": null, "source": { "name": "Microsoft Research \u2014 Lee et al. (2025), CHI 2025", "url": "https://www.microsoft.com/en-us/research/publication/the-impact-of-generative-ai-on-critical-thinking-self-reported-reductions-in-cognitive-effort-and-confidence-effects-from-a-survey-of-knowledge-workers/", "quote": "Higher confidence in GenAI is associated with less critical thinking, while higher self-confidence is associated with more critical thinking." }, "verification": { "status": "verified", "method": "full_quote", "coverage_pct": null, "fetch_mode": "live", "credibility": { "domain": "microsoft.com", "source_type": "unknown", "tier": 2, "flags": [], "note": "Unclassified domain \u2014 verify source authority manually" } }, "extraction": { "value": "verified", "value_in_quote": true, "quote_snippet": "Higher confidence in GenAI is associated with less critical thinking, while high" } }, "B3": { "type": "empirical", "label": "Harvard Gazette (2025): Harvard faculty experts warn AI use undercuts critical thinking", "sub_claim": null, "source": { "name": "Harvard Gazette (2025)", "url": "https://news.harvard.edu/gazette/story/2025/11/is-ai-dulling-our-minds/", "quote": "I am very worried about the effects of general-use LLMs on critical reasoning skills" }, "verification": { "status": "verified", "method": "full_quote", "coverage_pct": null, "fetch_mode": "live", "credibility": { "domain": "harvard.edu", "source_type": "academic", "tier": 4, "flags": [], "note": "Academic domain (.edu)" } }, "extraction": { "value": "verified", "value_in_quote": true, "quote_snippet": "I am very worried about the effects of general-use LLMs on critical reasoning sk" } }, "B4": { "type": "empirical", "label": "Jose et al. (2025, PMC): ChatGPT users scored 17% lower on concept understanding despite solving 48% more problems", "sub_claim": null, "source": { "name": "Jose et al. (2025), Frontiers \u2014 PMC", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC12036037/", "quote": "Excessive reliance may reduce cognitive engagement and long-term retention" }, "verification": { "status": "verified", "method": "full_quote", "coverage_pct": null, "fetch_mode": "live", "credibility": { "domain": "nih.gov", "source_type": "government", "tier": 5, "flags": [], "note": "Government domain (.gov)" } }, "extraction": { "value": "verified", "value_in_quote": true, "quote_snippet": "Excessive reliance may reduce cognitive engagement and long-term retention" } }, "A1": { "type": "computed", "label": "Verified source count meets threshold", "sub_claim": null, "method": "count(verified citations) = 4", "result": "4", "depends_on": [] } }, "cross_checks": [ { "description": "Multiple independent sources consulted", "n_sources_consulted": 4, "n_sources_verified": 4, "sources": { "gerlich_2025": "verified", "lee_chi_2025": "verified", "harvard_gazette_2025": "verified", "pmc_cognitive_paradox": "verified" }, "independence_note": "Sources are from different institutions and research teams: (1) SBS Swiss Business School via Phys.org, (2) Microsoft Research via CHI 2025, (3) Harvard University via Harvard Gazette, (4) Multiple Indian universities via PMC/Frontiers. No two sources share authors or datasets.", "fact_ids": [] } ], "adversarial_checks": [ { "question": "Do any studies show AI tools IMPROVE critical thinking or problem-solving?", "verification_performed": "Searched for 'AI tools improve critical thinking enhance problem solving evidence study 2025 2026'. Found that AI-powered classrooms can improve learning outcomes by 23-35% in STEM disciplines and language learning. Stanford research showed a 15% increase in scores for students using AI platforms. However, these gains are in knowledge acquisition, not in independent critical thinking or problem-solving ability. The PMC cognitive paradox paper itself notes that ChatGPT users solved 48% more problems but scored 17% lower on concept understanding \u2014 showing AI helps with task completion but may impair deeper cognitive engagement.", "finding": "AI tools can improve task performance and learning outcomes, but these benefits are distinct from critical thinking and independent problem-solving. The evidence consistently shows that while AI boosts productivity, it may simultaneously reduce the depth of cognitive engagement required for critical thinking.", "breaks_proof": false }, { "question": "Are the effects task-dependent rather than universal?", "verification_performed": "Searched for Microsoft CHI 2025 findings on task-dependent effects. The Lee et al. study found that for high-stakes tasks requiring accuracy, workers expend MORE effort in critical thinking with AI. For routine, low-stakes tasks under time pressure, they report LESS critical thinking effort. This shows the effect is moderated by task stakes and user confidence, not universal.", "finding": "The cognitive decline effect is moderated by task stakes, user confidence, and usage patterns. This does not break the proof because: (1) the claim is supported by the overall pattern across multiple studies, (2) the operator_note explicitly acknowledges this nuance, and (3) even task-dependent effects confirm that AI usage CAN and DOES reduce critical thinking under common conditions (routine tasks, high AI confidence). The proof documents this important qualification.", "breaks_proof": false }, { "question": "Has the key Gerlich (2025) study been retracted or significantly corrected?", "verification_performed": "Searched for 'Gerlich 2025 AI Tools in Society correction retraction'. Found a correction notice (Societies 2025, 15(9), 252) published September 2025. The correction addressed a duplicated table (Table 4 was a duplicate of Table 3). The author states the scientific conclusions are unaffected, and the correction was approved by the Academic Editor.", "finding": "The correction was minor (table duplication) and does not affect the study's findings or conclusions about the negative correlation between AI usage and critical thinking.", "breaks_proof": false } ], "verdict": { "value": "PROVED", "qualified": false, "qualifier": null, "reason": null }, "key_results": { "n_confirmed": 4, "threshold": 3, "operator": ">=", "claim_holds": true }, "generator": { "name": "proof-engine", "version": "1.2.0", "repo": "https://github.com/yaniv-golan/proof-engine", "generated_at": "2026-03-29" }, "proof_py_url": "/proofs/using-ai-tools-makes-humans-worse-at-critical-thin/proof.py", "citation": { "doi": "10.5281/zenodo.19455692", "concept_doi": "10.5281/zenodo.19454400", "url": "https://proofengine.info/proofs/using-ai-tools-makes-humans-worse-at-critical-thin/", "author": "Proof Engine", "cite_bib_url": "/proofs/using-ai-tools-makes-humans-worse-at-critical-thin/cite.bib", "cite_ris_url": "/proofs/using-ai-tools-makes-humans-worse-at-critical-thin/cite.ris" }, "depends_on": [] }