{ "format_version": 3, "claim_formal": { "subject": "Current AI systems (as of March 2026)", "property": "achievement of Artificial General Intelligence (AGI)", "operator": ">=", "operator_note": "This is a disproof. We search for authoritative sources that reject the claim that current AI systems have achieved AGI. AGI is interpreted using the most widely-cited frameworks: (1) Google DeepMind's 'Levels of AGI' paper (Morris et al., 2023), which classifies current frontier models as Level 1 ('Emerging') AGI \u2014 not yet 'Competent' (Level 2) on most cognitive tasks; (2) OpenAI's internal 5-level framework, which places current systems at Level 2 ('Reasoners') out of 5 levels needed for full AGI; (3) Expert survey consensus that AGI has not been achieved. The threshold of 3 independent authoritative sources rejecting the claim is conservative. If >= 3 verified sources explicitly state AGI has NOT been achieved, the claim is DISPROVED.", "threshold": 3, "proof_direction": "disprove" }, "claim_natural": "Current AI systems have already achieved Artificial General Intelligence (AGI).", "evidence": { "B1": { "type": "empirical", "label": "Google DeepMind 'Levels of AGI' framework \u2014 classifies current AI as Level 1 (Emerging)", "sub_claim": null, "source": { "name": "Google DeepMind (Morris et al., 2023) \u2014 Levels of AGI paper (arXiv:2311.02462)", "url": "https://arxiv.org/abs/2311.02462", "quote": "We propose a framework for classifying the capabilities and behavior of Artificial General Intelligence (AGI) models and their precursors. This framework introduces levels of AGI performance, generality, and autonomy, providing a common language to compare models, assess risks, and measure progress toward AGI." }, "verification": { "status": "verified", "method": "fragment", "coverage_pct": 81.4, "fetch_mode": "live", "credibility": { "domain": "arxiv.org", "source_type": "academic", "tier": 4, "flags": [], "note": "Known academic/scholarly publisher" } }, "extraction": { "value": "verified", "value_in_quote": true, "quote_snippet": "We propose a framework for classifying the capabilities and behavior of Artifici" } }, "B2": { "type": "empirical", "label": "Gary Marcus (NYU) \u2014 current AI is not AGI, conflates statistical approximation with intelligence", "sub_claim": null, "source": { "name": "Gary Marcus \u2014 'Rumors of AGI's arrival have been greatly exaggerated' (Substack)", "url": "https://garymarcus.substack.com/p/rumors-of-agis-arrival-have-been", "quote": "Current AI systems are powerful and increasingly useful tools, but they do not exhibit the flexible, self-directed competence that the original concept of artificial general intelligence was intended to capture." }, "verification": { "status": "verified", "method": "full_quote", "coverage_pct": null, "fetch_mode": "live", "credibility": { "domain": "substack.com", "source_type": "unknown", "tier": 2, "flags": [], "note": "Unclassified domain \u2014 verify source authority manually" } }, "extraction": { "value": "verified", "value_in_quote": true, "quote_snippet": "Current AI systems are powerful and increasingly useful tools, but they do not e" } }, "B3": { "type": "empirical", "label": "Expert analysis \u2014 AGI not achieved, current systems lack autonomous goals and transfer learning", "sub_claim": null, "source": { "name": "Cogni Down Under \u2014 'AGI Still Years Away' analysis (Medium)", "url": "https://medium.com/@cognidownunder/agi-still-years-away-despite-tech-leaders-bold-promises-for-2026-146c9780af65", "quote": "Current models lack autonomous goal formation. They respond brilliantly to prompts but never wonder what to explore on their own." }, "verification": { "status": "verified", "method": "full_quote", "coverage_pct": null, "fetch_mode": "live", "credibility": { "domain": "medium.com", "source_type": "unknown", "tier": 2, "flags": [], "note": "Unclassified domain \u2014 verify source authority manually" } }, "extraction": { "value": "verified", "value_in_quote": true, "quote_snippet": "Current models lack autonomous goal formation. They respond brilliantly to promp" } }, "B4": { "type": "empirical", "label": "Tim Dettmers (UW) \u2014 AGI will not happen due to physical computation limits", "sub_claim": null, "source": { "name": "Tim Dettmers (University of Washington) \u2014 'Why AGI Will Not Happen' (2025)", "url": "https://timdettmers.com/2025/12/10/why-agi-will-not-happen/", "quote": "For linear improvements, we previously had exponential growth as GPUs which canceled out the exponential resource requirements of scaling. This is no longer true. In other words, previously we invested roughly linear costs to get linear payoff, but now it has turned to exponential costs." }, "verification": { "status": "verified", "method": "full_quote", "coverage_pct": null, "fetch_mode": "live", "credibility": { "domain": "timdettmers.com", "source_type": "unknown", "tier": 2, "flags": [], "note": "Unclassified domain \u2014 verify source authority manually" } }, "extraction": { "value": "verified", "value_in_quote": true, "quote_snippet": "For linear improvements, we previously had exponential growth as GPUs which canc" } }, "A1": { "type": "computed", "label": "Verified source count meeting disproof 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": { "deepmind_levels": "verified", "gary_marcus": "verified", "cogni_analysis": "verified", "tim_dettmers": "verified" }, "independence_note": "Sources are from different institutions and individuals: Google DeepMind (academic research lab), Gary Marcus (NYU professor/independent researcher), Cogni Down Under (independent AI analysis), Tim Dettmers (University of Washington). Each reaches the same conclusion via different reasoning: DeepMind via formal taxonomy, Marcus via philosophy of mind, Cogni via capability analysis, Dettmers via physical limits.", "fact_ids": [] } ], "adversarial_checks": [ { "question": "Has any credible AI researcher or organization officially declared AGI achieved?", "verification_performed": "Searched for 'AGI achieved 2026 claims'. Found that Nvidia CEO Jensen Huang stated 'I think we've achieved AGI' (March 2026), but this was widely criticized by the research community. His definition relied narrowly on AI passing human exams, which experts note tests narrow competencies, not general intelligence. OpenAI, Google DeepMind, and Anthropic have NOT claimed AGI achievement. OpenAI places current systems at Level 2 of 5 on their internal AGI framework.", "finding": "Jensen Huang's claim is the only major industry figure to declare AGI achieved. His claim was immediately challenged by researchers who note it conflates benchmark performance with general intelligence. No major AI research lab has endorsed the claim. 76% of 475 AI researchers surveyed by AAAI said scaling current AI is unlikely to result in AGI.", "breaks_proof": false }, { "question": "Do current AI systems pass any widely-accepted AGI benchmark or test?", "verification_performed": "Searched for 'AGI benchmark test passed 2026'. Found that while current LLMs pass many standardized exams (bar exam, medical licensing, math olympiads), experts argue these test narrow competencies. DeepMind's 2026 cognitive framework identifies 10 key cognitive abilities for AGI, and notes the largest gaps are in learning, metacognition, attention, executive functions, and social cognition \u2014 areas where current systems fundamentally underperform.", "finding": "No widely-accepted AGI benchmark has been passed. Current systems show 'jagged intelligence' \u2014 winning math olympiad gold medals but failing elementary problems. DeepMind's framework shows large gaps in 5 of 10 cognitive abilities needed for AGI.", "breaks_proof": false }, { "question": "Is there expert consensus that AGI timelines are imminent (already here or within 1 year)?", "verification_performed": "Searched for 'AGI timeline expert survey 2025 2026'. Found that forecasters average a 25% chance of AGI by 2029 and 50% by 2033 (as of Feb 2026). Stanford HAI experts stated 'There will be no AGI this year' for 2026. A 2023 survey of AI researchers predicted AGI around 2040 on average.", "finding": "Expert consensus places AGI arrival well into the future. Even optimistic forecasters give only 25% probability by 2029. No mainstream expert survey claims AGI is already here.", "breaks_proof": false } ], "verdict": { "value": "DISPROVED", "qualified": false, "qualifier": null, "reason": null }, "key_results": { "n_confirmed": 4, "threshold": 3, "operator": ">=", "claim_holds": true }, "generator": { "name": "proof-engine", "version": "1.7.0", "repo": "https://github.com/yaniv-golan/proof-engine", "generated_at": "2026-04-06" }, "proof_py_url": "/proofs/current-ai-systems-have-already-achieved-artificia/proof.py", "citation": { "doi": "10.5281/zenodo.19489830", "concept_doi": "10.5281/zenodo.19489829", "url": "https://proofengine.info/proofs/current-ai-systems-have-already-achieved-artificia/", "author": "Proof Engine", "cite_bib_url": "/proofs/current-ai-systems-have-already-achieved-artificia/cite.bib", "cite_ris_url": "/proofs/current-ai-systems-have-already-achieved-artificia/cite.ris" }, "depends_on": [] }