{
  "name": "AGI timeline predictions: when experts, surveys and forecasters say AGI will happen",
  "description": "AGI timeline predictions from lab leaders, researchers, AI Impacts surveys, Metaculus and AI 2027, each dated and sourced. Free CSV and JSON download.",
  "url": "https://agitracker.io/data/agi-timeline-predictions/",
  "license": "https://creativecommons.org/licenses/by/4.0/",
  "publisher": "agitracker.io",
  "updated": "2026-10-29",
  "columns": [
    {
      "key": "person",
      "label": "Person or source"
    },
    {
      "key": "org",
      "label": "Role or organization",
      "note": "Role at the time of the statement"
    },
    {
      "key": "predicted",
      "label": "Predicted year or range",
      "note": "As stated, or converted from \"N years\" to calendar years"
    },
    {
      "key": "midpoint",
      "label": "Midpoint year",
      "type": "number",
      "note": "Single year or midpoint of a stated range; blank when no year was given"
    },
    {
      "key": "prediction",
      "label": "What was predicted",
      "note": "Paraphrase keeping the source's milestone wording"
    },
    {
      "key": "date",
      "label": "Date of statement",
      "type": "date"
    },
    {
      "key": "type",
      "label": "Type",
      "note": "Individual statement, expert survey, forecasting platform, scenario, or forecasting model"
    }
  ],
  "rows": [
    {
      "person": "Sam Altman",
      "org": "CEO, OpenAI",
      "predicted": "No year given (\"a few thousand days\" from Sept 2024)",
      "midpoint": null,
      "prediction": "Superintelligence possibly within a few thousand days, though it may take longer.",
      "date": "2024-09-23",
      "type": "Individual statement",
      "source": {
        "title": "Sam Altman: The Intelligence Age",
        "url": "https://ia.samaltman.com/"
      }
    },
    {
      "person": "Sam Altman",
      "org": "CEO, OpenAI",
      "predicted": "2025-2029",
      "midpoint": 2027,
      "prediction": "AGI will probably be developed during the U.S. presidential term that began in January 2025.",
      "date": "2025-01-05",
      "type": "Individual statement",
      "source": {
        "title": "VentureBeat on Altman's Bloomberg Businessweek interview",
        "url": "https://venturebeat.com/ai/openai-begins-2025-with-massive-hype-for-agi-superintelligence"
      }
    },
    {
      "person": "Sam Altman",
      "org": "CEO, OpenAI",
      "predicted": "2025",
      "midpoint": 2025,
      "prediction": "OpenAI is confident it knows how to build AGI \"as we have traditionally understood it\"; first AI agents may \"join the workforce\" in 2025.",
      "date": "2025-01-05",
      "type": "Individual statement",
      "source": {
        "title": "Sam Altman: Reflections",
        "url": "https://blog.samaltman.com/reflections"
      }
    },
    {
      "person": "Sam Altman",
      "org": "CEO, OpenAI",
      "predicted": "2026",
      "midpoint": 2026,
      "prediction": "Systems that can figure out novel insights likely in 2026; robots doing real-world tasks may follow in 2027.",
      "date": "2025-06-10",
      "type": "Individual statement",
      "source": {
        "title": "Sam Altman: The Gentle Singularity",
        "url": "https://blog.samaltman.com/the-gentle-singularity"
      }
    },
    {
      "person": "Sam Altman",
      "org": "CEO, OpenAI",
      "predicted": "By 2030",
      "midpoint": null,
      "prediction": "Would be very surprised if there were no models that are extraordinarily capable and do things humans cannot do by 2030 (asked about superintelligence).",
      "date": "2025-09-25",
      "type": "Individual statement",
      "source": {
        "title": "Politico: Altman interview with Axel Springer",
        "url": "https://www.politico.com/news/magazine/2025/09/25/sam-altman-ai-interview-axel-springer-00580997"
      }
    },
    {
      "person": "Sam Altman",
      "org": "CEO, OpenAI",
      "predicted": "No year given",
      "midpoint": null,
      "prediction": "AGI is \"pretty close\" and superintelligence \"not that far off\"; expects a faster takeoff than he once thought.",
      "date": "2026-02",
      "type": "Individual statement",
      "source": {
        "title": "The Decoder: Altman at Express Adda",
        "url": "https://the-decoder.com/altman-warns-the-world-is-not-prepared-as-openai-accelerates-research-using-its-own-ai/"
      }
    },
    {
      "person": "Dario Amodei",
      "org": "CEO, Anthropic",
      "predicted": "2025-2026",
      "midpoint": 2025.5,
      "prediction": "A model that talks like a generally well-educated human could arrive in two or three years.",
      "date": "2023-08-08",
      "type": "Individual statement",
      "source": {
        "title": "Dwarkesh Podcast: Dario Amodei (transcript)",
        "url": "https://www.dwarkesh.com/p/dario-amodei"
      }
    },
    {
      "person": "Dario Amodei",
      "org": "CEO, Anthropic",
      "predicted": "As early as 2026",
      "midpoint": null,
      "prediction": "\"Powerful AI\" (smarter than a Nobel Prize winner across most fields) could come as early as 2026, or take much longer.",
      "date": "2024-10",
      "type": "Individual statement",
      "source": {
        "title": "Dario Amodei: Machines of Loving Grace",
        "url": "https://www.darioamodei.com/essay/machines-of-loving-grace"
      }
    },
    {
      "person": "Dario Amodei",
      "org": "CEO, Anthropic",
      "predicted": "2026-2027",
      "midpoint": 2026.5,
      "prediction": "Extrapolating capability trends suggests reaching powerful, human-expert-level AI by 2026 or 2027, with many possible derailers.",
      "date": "2024-11-11",
      "type": "Individual statement",
      "source": {
        "title": "Lex Fridman Podcast #452: Dario Amodei (transcript)",
        "url": "https://lexfridman.com/dario-amodei-transcript"
      }
    },
    {
      "person": "Dario Amodei",
      "org": "CEO, Anthropic",
      "predicted": "About 2027",
      "midpoint": 2027,
      "prediction": "AI systems better than humans at almost everything; unsure whether it will be 2027, but not a whole lot longer.",
      "date": "2025-01-21",
      "type": "Individual statement",
      "source": {
        "title": "Ars Technica: Amodei at Davos 2025",
        "url": "https://arstechnica.com/ai/2025/01/anthropic-chief-says-ai-could-surpass-almost-all-humans-at-almost-everything-shortly-after-2027/"
      }
    },
    {
      "person": "Anthropic (company)",
      "org": "AI lab",
      "predicted": "Late 2026 to early 2027",
      "midpoint": 2026.5,
      "prediction": "Expects \"powerful AI systems\" with Nobel-level intellectual capabilities across most disciplines to emerge in late 2026 or early 2027.",
      "date": "2025-03-06",
      "type": "Individual statement",
      "source": {
        "title": "Anthropic: Recommendations to OSTP for the U.S. AI Action Plan",
        "url": "https://www.anthropic.com/news/anthropic-s-recommendations-ostp-u-s-ai-action-plan"
      }
    },
    {
      "person": "Dario Amodei",
      "org": "CEO, Anthropic",
      "predicted": "2027-2028",
      "midpoint": 2027.5,
      "prediction": "AI models doing \"Nobel-level\" scientific research in multiple fields within two years.",
      "date": "2026-01",
      "type": "Individual statement",
      "source": {
        "title": "Fortune: AI leaders at Davos 2026",
        "url": "https://fortune.com/2026/01/23/deepmind-demis-hassabis-anthropic-dario-amodei-yann-lecun-ai-davos/"
      }
    },
    {
      "person": "Demis Hassabis",
      "org": "CEO, Google DeepMind",
      "predicted": "2030-2035",
      "midpoint": 2032.5,
      "prediction": "AGI, a system with all the complex capabilities humans have, will start to emerge in the next five to 10 years.",
      "date": "2025-03-17",
      "type": "Individual statement",
      "source": {
        "title": "CNBC: Hassabis on AGI in five to 10 years",
        "url": "https://www.cnbc.com/2025/03/17/human-level-ai-will-be-here-in-5-to-10-years-deepmind-ceo-says.html"
      }
    },
    {
      "person": "Demis Hassabis",
      "org": "CEO, Google DeepMind",
      "predicted": "Just after 2030",
      "midpoint": 2030,
      "prediction": "Asked whether AGI arrives before or after 2030, chose just after.",
      "date": "2025-05-20",
      "type": "Individual statement",
      "source": {
        "title": "Axios: Hassabis and Brin at Google I/O 2025",
        "url": "https://www.axios.com/2025/05/21/google-sergey-brin-demis-hassabis-agi-2030"
      }
    },
    {
      "person": "Sergey Brin",
      "org": "Co-founder, Google",
      "predicted": "Just before 2030",
      "midpoint": 2030,
      "prediction": "Asked whether AGI arrives before or after 2030, chose just before.",
      "date": "2025-05-20",
      "type": "Individual statement",
      "source": {
        "title": "Axios: Hassabis and Brin at Google I/O 2025",
        "url": "https://www.axios.com/2025/05/21/google-sergey-brin-demis-hassabis-agi-2030"
      }
    },
    {
      "person": "Sundar Pichai",
      "org": "CEO, Google and Alphabet",
      "predicted": "Slightly after 2030",
      "midpoint": 2030,
      "prediction": "Expects AI to fall just short of AGI by 2030, taking a bit longer, while progress will be dramatic.",
      "date": "2025-06-05",
      "type": "Individual statement",
      "source": {
        "title": "Lex Fridman Podcast #471: Sundar Pichai (transcript)",
        "url": "https://lexfridman.com/sundar-pichai-transcript/"
      }
    },
    {
      "person": "Demis Hassabis",
      "org": "CEO, Google DeepMind",
      "predicted": "By 2030 (50%)",
      "midpoint": 2030,
      "prediction": "A 50% chance AGI is achieved within the decade, likely needing one or two more breakthroughs.",
      "date": "2026-01",
      "type": "Individual statement",
      "source": {
        "title": "Fortune: AI leaders at Davos 2026",
        "url": "https://fortune.com/2026/01/23/deepmind-demis-hassabis-anthropic-dario-amodei-yann-lecun-ai-davos/"
      }
    },
    {
      "person": "Shane Legg",
      "org": "Co-founder, DeepMind",
      "predicted": "2028 (mean)",
      "midpoint": 2028,
      "prediction": "Human-level AGI: log-normal distribution with a mean of 2028 and a mode of 2025.",
      "date": "2011-12-31",
      "type": "Individual statement",
      "source": {
        "title": "Shane Legg: Goodbye 2011, hello 2012",
        "url": "http://www.vetta.org/2011/12/goodbye-2011-hello-2012/"
      }
    },
    {
      "person": "Shane Legg",
      "org": "Co-founder and Chief AGI Scientist, Google DeepMind",
      "predicted": "2028 (50%)",
      "midpoint": 2028,
      "prediction": "A 50% chance of AGI by 2028.",
      "date": "2023-10-26",
      "type": "Individual statement",
      "source": {
        "title": "Dwarkesh Podcast: Shane Legg (transcript)",
        "url": "https://www.dwarkesh.com/p/shane-legg"
      }
    },
    {
      "person": "Elon Musk",
      "org": "CEO, Tesla and xAI",
      "predicted": "2025-2026",
      "midpoint": 2025.5,
      "prediction": "If AGI means smarter than the smartest human, probably next year, within two years.",
      "date": "2024-04-08",
      "type": "Individual statement",
      "source": {
        "title": "Ars Technica: Musk on AI smarter than any human",
        "url": "https://arstechnica.com/information-technology/2024/04/elon-musk-ai-will-be-smarter-than-any-human-around-the-end-of-next-year/"
      }
    },
    {
      "person": "Elon Musk",
      "org": "CEO, Tesla and xAI",
      "predicted": "2026-2027",
      "midpoint": 2026.5,
      "prediction": "AI smarter than any human by the end of 2026, no later than 2027.",
      "date": "2026-01-22",
      "type": "Individual statement",
      "source": {
        "title": "Decrypt: Musk at Davos 2026",
        "url": "https://decrypt.co/355532/elon-musk-predicts-ai-outpace-human-intelligence-this-year"
      }
    },
    {
      "person": "Elon Musk",
      "org": "CEO, Tesla and xAI",
      "predicted": "2030-2031",
      "midpoint": 2030.5,
      "prediction": "AI smarter than all of humanity collectively, probably by 2030 or 2031.",
      "date": "2026-01-22",
      "type": "Individual statement",
      "source": {
        "title": "Decrypt: Musk at Davos 2026",
        "url": "https://decrypt.co/355532/elon-musk-predicts-ai-outpace-human-intelligence-this-year"
      }
    },
    {
      "person": "Mark Zuckerberg",
      "org": "CEO, Meta",
      "predicted": "No year given (this decade decisive)",
      "midpoint": null,
      "prediction": "Developing superintelligence is \"now in sight\"; the rest of this decade is likely the decisive period.",
      "date": "2025-07-30",
      "type": "Individual statement",
      "source": {
        "title": "Meta: Personal Superintelligence (Mark Zuckerberg)",
        "url": "https://www.meta.com/superintelligence/"
      }
    },
    {
      "person": "Mustafa Suleyman",
      "org": "CEO, Microsoft AI",
      "predicted": "2029-2031",
      "midpoint": 2030,
      "prediction": "AGI is not possible on current chips but plausible within the next five to seven years (two to five hardware generations).",
      "date": "2024-12-09",
      "type": "Individual statement",
      "source": {
        "title": "The Verge Decoder: Mustafa Suleyman interview",
        "url": "https://www.theverge.com/24314821/microsoft-ai-ceo-mustafa-suleyman-google-deepmind-openai-inflection-agi-decoder-podcast"
      }
    },
    {
      "person": "Ilya Sutskever",
      "org": "Co-founder, Safe Superintelligence",
      "predicted": "2030-2045",
      "midpoint": 2037.5,
      "prediction": "A system that learns as well as a human and then becomes superhuman: \"like 5 to 20\" years.",
      "date": "2025-11-25",
      "type": "Individual statement",
      "source": {
        "title": "Dwarkesh Podcast: Ilya Sutskever (transcript)",
        "url": "https://www.dwarkesh.com/p/ilya-sutskever-2"
      }
    },
    {
      "person": "Jensen Huang",
      "org": "CEO, Nvidia",
      "predicted": "By 2029",
      "midpoint": null,
      "prediction": "If AGI is defined as doing well on a specific set of tests (bar exam, logic tests and so on), within five years.",
      "date": "2024-03-19",
      "type": "Individual statement",
      "source": {
        "title": "TechCrunch: Jensen Huang on AGI at GTC 2024",
        "url": "https://techcrunch.com/2024/03/19/agi-and-hallucinations/"
      }
    },
    {
      "person": "Masayoshi Son",
      "org": "CEO, SoftBank",
      "predicted": "About 2035",
      "midpoint": 2035,
      "prediction": "Artificial superintelligence 10,000 times smarter than a human brain within about a decade.",
      "date": "2024-10-29",
      "type": "Individual statement",
      "source": {
        "title": "The Register: Masayoshi Son at FII 2024",
        "url": "https://www.theregister.com/software/2024/10/29/softbank-ceo-says-super-ai-will-be-here-in-2035-and-cost-9t/354098"
      }
    },
    {
      "person": "Geoffrey Hinton",
      "org": "Turing Award and Nobel laureate",
      "predicted": "By 2043",
      "midpoint": null,
      "prediction": "General-purpose AI may be \"20 years or less\" away, down from his earlier 20 to 50 years.",
      "date": "2023-03-25",
      "type": "Individual statement",
      "source": {
        "title": "CBS News: Geoffrey Hinton interview, March 2023",
        "url": "https://www.cbsnews.com/news/godfather-of-artificial-intelligence-weighs-in-on-the-past-and-potential-of-artificial-intelligence/"
      }
    },
    {
      "person": "Yoshua Bengio",
      "org": "Professor, Mila / Universite de Montreal",
      "predicted": "2028-2043 (90% confidence)",
      "midpoint": 2035.5,
      "prediction": "Human-level AI in 5 to 20 years with 90% confidence, revised from decades to centuries.",
      "date": "2023-08-12",
      "type": "Individual statement",
      "source": {
        "title": "Yoshua Bengio: Personal and psychological dimensions of AI researchers confronting AI catastrophic risks",
        "url": "https://yoshuabengio.org/en/blog/personal-and-psychological-dimensions-ai-researchers-confronting-ai-catastrophic-risks"
      }
    },
    {
      "person": "Geoffrey Hinton",
      "org": "Turing Award and Nobel laureate",
      "predicted": "By about 2035",
      "midpoint": null,
      "prediction": "A good chance AI that surpasses humans in every domain arrives in 10 years or less (a year earlier he said 5 to 20).",
      "date": "2025-04",
      "type": "Individual statement",
      "source": {
        "title": "Business Insider: Hinton on superintelligence timeline",
        "url": "https://www.businessinsider.com/ai-godfather-geoffrey-hinton-superintelligence-risk-takeover-2025-4"
      }
    },
    {
      "person": "Geoffrey Hinton",
      "org": "Turing Award and Nobel laureate",
      "predicted": "2030-2045",
      "midpoint": 2037.5,
      "prediction": "AGI: \"a reasonable bet is sometime between five and 20 years.\"",
      "date": "2025-08",
      "type": "Individual statement",
      "source": {
        "title": "CNN: Geoffrey Hinton at Ai4, August 2025",
        "url": "https://edition.cnn.com/2025/08/13/tech/ai-geoffrey-hinton"
      }
    },
    {
      "person": "Yann LeCun",
      "org": "Chief AI Scientist, Meta (at the time)",
      "predicted": "Years to a decade or more",
      "midpoint": null,
      "prediction": "Human-level AI could be \"years to decades\" away; getting world models to work will take years, if not a decade.",
      "date": "2024-10-16",
      "type": "Individual statement",
      "source": {
        "title": "TechCrunch: LeCun on world models and human-level AI",
        "url": "https://techcrunch.com/2024/10/16/metas-ai-chief-says-world-models-are-key-to-human-level-ai-but-it-might-be-10-years-out/"
      }
    },
    {
      "person": "Ray Kurzweil",
      "org": "Futurist, author",
      "predicted": "2029",
      "midpoint": 2029,
      "prediction": "The Singularity Is Near (2005): computers reach human-level intelligence by 2029; the Singularity around 2045.",
      "date": "2005",
      "type": "Individual statement",
      "source": {
        "title": "The Guardian: Ray Kurzweil interview, June 2024",
        "url": "https://www.theguardian.com/technology/article/2024/jun/29/ray-kurzweil-google-ai-the-singularity-is-nearer"
      }
    },
    {
      "person": "Ray Kurzweil",
      "org": "Futurist, author",
      "predicted": "2029",
      "midpoint": 2029,
      "prediction": "Reaffirms 2029 for both human-level intelligence and AGI, and 2045 for the Singularity.",
      "date": "2024-06-29",
      "type": "Individual statement",
      "source": {
        "title": "The Guardian: Ray Kurzweil interview, June 2024",
        "url": "https://www.theguardian.com/technology/article/2024/jun/29/ray-kurzweil-google-ai-the-singularity-is-nearer"
      }
    },
    {
      "person": "Andrej Karpathy",
      "org": "Founder, Eureka Labs; OpenAI co-founder",
      "predicted": "About 2035",
      "midpoint": 2035,
      "prediction": "AGI (agents that work like an employee or intern you would hire) is still about a decade away.",
      "date": "2025-10-17",
      "type": "Individual statement",
      "source": {
        "title": "Dwarkesh Podcast: Andrej Karpathy, AGI is still a decade away",
        "url": "https://www.dwarkesh.com/p/andrej-karpathy"
      }
    },
    {
      "person": "Leopold Aschenbrenner",
      "org": "Former OpenAI researcher",
      "predicted": "2027",
      "midpoint": 2027,
      "prediction": "AGI by 2027 is \"strikingly plausible\" based on trends in compute and algorithmic efficiency.",
      "date": "2024-06",
      "type": "Individual statement",
      "source": {
        "title": "Leopold Aschenbrenner: Situational Awareness",
        "url": "https://situational-awareness.ai/"
      }
    },
    {
      "person": "Ajeya Cotra",
      "org": "Open Philanthropy",
      "predicted": "~2050 (median)",
      "midpoint": 2050,
      "prediction": "Median of about 2050 for transformative AI from the 2020 biological anchors report.",
      "date": "2020",
      "type": "Individual statement",
      "source": {
        "title": "Ajeya Cotra: Two-year update on my personal AI timelines",
        "url": "https://www.alignmentforum.org/posts/AfH2oPHCApdKicM4m/two-year-update-on-my-personal-ai-timelines"
      }
    },
    {
      "person": "Ajeya Cotra",
      "org": "Open Philanthropy",
      "predicted": "~2040 (median)",
      "midpoint": 2040,
      "prediction": "Updated median of about 2040 for transformative AI, roughly 10 years earlier than in 2020.",
      "date": "2022-08-02",
      "type": "Individual statement",
      "source": {
        "title": "Ajeya Cotra: Two-year update on my personal AI timelines",
        "url": "https://www.alignmentforum.org/posts/AfH2oPHCApdKicM4m/two-year-update-on-my-personal-ai-timelines"
      }
    },
    {
      "person": "Daniel Kokotajlo",
      "org": "Forecaster (later AI Futures Project)",
      "predicted": "2029 (median)",
      "midpoint": 2029,
      "prediction": "AGI median of about 2029.",
      "date": "2021-08",
      "type": "Individual statement",
      "source": {
        "title": "AI Futures Project: How our AI timelines forecasts have changed since AI 2027",
        "url": "https://blog.aifutures.org/p/clarifying-how-our-ai-timelines-forecasts"
      }
    },
    {
      "person": "Daniel Kokotajlo",
      "org": "OpenAI (at the time)",
      "predicted": "2027 (median)",
      "midpoint": 2027,
      "prediction": "AGI median dropped to 2027 after joining OpenAI.",
      "date": "2022-12",
      "type": "Individual statement",
      "source": {
        "title": "AI Futures Project: How our AI timelines forecasts have changed since AI 2027",
        "url": "https://blog.aifutures.org/p/clarifying-how-our-ai-timelines-forecasts"
      }
    },
    {
      "person": "Daniel Kokotajlo",
      "org": "Executive Director, AI Futures Project",
      "predicted": "2028 (median)",
      "midpoint": 2028,
      "prediction": "AGI median moved to 2028, up from 2027.",
      "date": "2025-02",
      "type": "Individual statement",
      "source": {
        "title": "AI Futures Project: How our AI timelines forecasts have changed since AI 2027",
        "url": "https://blog.aifutures.org/p/clarifying-how-our-ai-timelines-forecasts"
      }
    },
    {
      "person": "Daniel Kokotajlo",
      "org": "Executive Director, AI Futures Project",
      "predicted": "Around 2030",
      "midpoint": 2030,
      "prediction": "Says \"around 2030, lots of uncertainty\" for AGI; progress somewhat slower than the AI 2027 scenario.",
      "date": "2025-11",
      "type": "Individual statement",
      "source": {
        "title": "AI Futures Project: How our AI timelines forecasts have changed since AI 2027",
        "url": "https://blog.aifutures.org/p/clarifying-how-our-ai-timelines-forecasts"
      }
    },
    {
      "person": "Daniel Kokotajlo",
      "org": "Executive Director, AI Futures Project",
      "predicted": "Dec 2030 (median)",
      "midpoint": 2030,
      "prediction": "All-things-considered median for TED-AI (top-expert-dominating AI, the project's AGI definition).",
      "date": "2026-01",
      "type": "Individual statement",
      "source": {
        "title": "AI Futures Project: How our AI timelines forecasts have changed since AI 2027",
        "url": "https://blog.aifutures.org/p/clarifying-how-our-ai-timelines-forecasts"
      }
    },
    {
      "person": "Daniel Kokotajlo",
      "org": "Executive Director, AI Futures Project",
      "predicted": "Mid-2028 (median)",
      "midpoint": 2028,
      "prediction": "Automated Coder (full automation of coding) median moved from late 2029 to mid-2028.",
      "date": "2026-04-02",
      "type": "Individual statement",
      "source": {
        "title": "AI Futures Project: Q1 2026 Timelines Update",
        "url": "https://blog.aifutures.org/p/q1-2026-timelines-update"
      }
    },
    {
      "person": "Eli Lifland",
      "org": "AI Futures Project",
      "predicted": "2060 (median)",
      "midpoint": 2060,
      "prediction": "Transformative AI timelines similar to about a 50% chance by 2060.",
      "date": "2021",
      "type": "Individual statement",
      "source": {
        "title": "AI Futures Project: How our AI timelines forecasts have changed since AI 2027",
        "url": "https://blog.aifutures.org/p/clarifying-how-our-ai-timelines-forecasts"
      }
    },
    {
      "person": "Eli Lifland",
      "org": "AI Futures Project",
      "predicted": "2050 (median)",
      "midpoint": 2050,
      "prediction": "AGI median of about 2050 after underestimating progress on MATH and MMLU.",
      "date": "2022-07",
      "type": "Individual statement",
      "source": {
        "title": "AI Futures Project: How our AI timelines forecasts have changed since AI 2027",
        "url": "https://blog.aifutures.org/p/clarifying-how-our-ai-timelines-forecasts"
      }
    },
    {
      "person": "Eli Lifland",
      "org": "AI Futures Project",
      "predicted": "2038 (median)",
      "midpoint": 2038,
      "prediction": "AGI median of 2038.",
      "date": "2024-01",
      "type": "Individual statement",
      "source": {
        "title": "AI Futures Project: How our AI timelines forecasts have changed since AI 2027",
        "url": "https://blog.aifutures.org/p/clarifying-how-our-ai-timelines-forecasts"
      }
    },
    {
      "person": "Eli Lifland",
      "org": "AI Futures Project",
      "predicted": "2031 (median)",
      "midpoint": 2031,
      "prediction": "AGI (TED-AI) median of 2031 at the time AI 2027 was published.",
      "date": "2025-04",
      "type": "Individual statement",
      "source": {
        "title": "AI Futures Project: How our AI timelines forecasts have changed since AI 2027",
        "url": "https://blog.aifutures.org/p/clarifying-how-our-ai-timelines-forecasts"
      }
    },
    {
      "person": "Eli Lifland",
      "org": "AI Futures Project",
      "predicted": "Jan 2035 (median)",
      "midpoint": 2035,
      "prediction": "All-things-considered TED-AI (AGI) median.",
      "date": "2026-01",
      "type": "Individual statement",
      "source": {
        "title": "AI Futures Project: How our AI timelines forecasts have changed since AI 2027",
        "url": "https://blog.aifutures.org/p/clarifying-how-our-ai-timelines-forecasts"
      }
    },
    {
      "person": "AI 2027 scenario",
      "org": "AI Futures Project (Kokotajlo, Alexander, Larsen, Lifland, Dean)",
      "predicted": "March 2027",
      "midpoint": 2027,
      "prediction": "Superhuman coder (SC) achieved in the scenario's racing ending.",
      "date": "2025-04-03",
      "type": "Scenario",
      "source": {
        "title": "AI 2027 (AI Futures Project)",
        "url": "https://ai-2027.com/"
      }
    },
    {
      "person": "AI 2027 scenario",
      "org": "AI Futures Project",
      "predicted": "August 2027",
      "midpoint": 2027,
      "prediction": "Superhuman AI researcher (SAR) achieved in the scenario.",
      "date": "2025-04-03",
      "type": "Scenario",
      "source": {
        "title": "AI 2027 (AI Futures Project)",
        "url": "https://ai-2027.com/"
      }
    },
    {
      "person": "AI 2027 scenario",
      "org": "AI Futures Project",
      "predicted": "December 2027",
      "midpoint": 2027,
      "prediction": "Artificial superintelligence (ASI), much better than the best human at every cognitive task, in the scenario.",
      "date": "2025-04-03",
      "type": "Scenario",
      "source": {
        "title": "AI 2027 (AI Futures Project)",
        "url": "https://ai-2027.com/"
      }
    },
    {
      "person": "AI 2027 timelines forecast (Lifland)",
      "org": "AI Futures Project",
      "predicted": "2030 (median; 80% CI 2026 to >2050)",
      "midpoint": 2030,
      "prediction": "All-things-considered median for a superhuman coder.",
      "date": "2025-04",
      "type": "Scenario",
      "source": {
        "title": "AI 2027: Timelines Forecast supplement",
        "url": "https://ai-2027.com/research/timelines-forecast"
      }
    },
    {
      "person": "AI 2027 timelines forecast (Jurkovic)",
      "org": "AI Futures Project",
      "predicted": "2028 (median; 80% CI 2026-2040)",
      "midpoint": 2028,
      "prediction": "All-things-considered median for a superhuman coder.",
      "date": "2025-04",
      "type": "Scenario",
      "source": {
        "title": "AI 2027: Timelines Forecast supplement",
        "url": "https://ai-2027.com/research/timelines-forecast"
      }
    },
    {
      "person": "AI 2027 timelines forecast (FutureSearch)",
      "org": "FutureSearch professional forecasters (n=3)",
      "predicted": "2033 (median; 80% CI 2027 to >2050)",
      "midpoint": 2033,
      "prediction": "Aggregate all-things-considered median for a superhuman coder.",
      "date": "2025-04",
      "type": "Scenario",
      "source": {
        "title": "AI 2027: Timelines Forecast supplement",
        "url": "https://ai-2027.com/research/timelines-forecast"
      }
    },
    {
      "person": "Muller and Bostrom survey",
      "org": "Survey of four expert groups, 2012/2013",
      "predicted": "2040-2050 (50%)",
      "midpoint": 2045,
      "prediction": "Median respondent gave a 50% chance of high-level machine intelligence around 2040-2050 and 90% by 2075.",
      "date": "2013",
      "type": "Expert survey",
      "source": {
        "title": "Muller and Bostrom: Future progress in artificial intelligence (arXiv)",
        "url": "https://arxiv.org/abs/2508.11681"
      }
    },
    {
      "person": "ESPAI 2016 (AI Impacts)",
      "org": "352 researchers who published at NIPS and ICML 2015",
      "predicted": "2061 (50%)",
      "midpoint": 2061,
      "prediction": "Aggregate 50% chance of HLMI (machines outperform humans at every task) within 45 years of 2016.",
      "date": "2016",
      "type": "Expert survey",
      "source": {
        "title": "Grace et al.: When Will AI Exceed Human Performance? (arXiv)",
        "url": "https://arxiv.org/abs/1705.08807"
      }
    },
    {
      "person": "ESPAI 2016 (AI Impacts)",
      "org": "352 researchers who published at NIPS and ICML 2015",
      "predicted": "About 120 years after 2016 (50%)",
      "midpoint": null,
      "prediction": "Aggregate 50% chance of full automation of all human jobs (FAOL) in about 120 years.",
      "date": "2016",
      "type": "Expert survey",
      "source": {
        "title": "Grace et al.: When Will AI Exceed Human Performance? (arXiv)",
        "url": "https://arxiv.org/abs/1705.08807"
      }
    },
    {
      "person": "ESPAI 2022 (AI Impacts)",
      "org": "738 researchers who published at NeurIPS and ICML 2021",
      "predicted": "2059 (50%)",
      "midpoint": 2059,
      "prediction": "Aggregate 50% chance of HLMI by 2059, 37 years after the survey (later reported as 2060).",
      "date": "2022",
      "type": "Expert survey",
      "source": {
        "title": "AI Impacts: 2022 Expert Survey on Progress in AI",
        "url": "https://aiimpacts.org/2022-expert-survey-on-progress-in-ai/"
      }
    },
    {
      "person": "ESPAI 2022 (AI Impacts)",
      "org": "738 researchers who published at NeurIPS and ICML 2021",
      "predicted": "2164 (50%)",
      "midpoint": 2164,
      "prediction": "Aggregate 50% chance of full automation of labor (FAOL) by 2164.",
      "date": "2022",
      "type": "Expert survey",
      "source": {
        "title": "Grace et al.: Thousands of AI Authors on the Future of AI (arXiv)",
        "url": "https://arxiv.org/abs/2401.02843"
      }
    },
    {
      "person": "ESPAI 2023 (AI Impacts)",
      "org": "2,778 researchers at six top AI venues",
      "predicted": "2047 (50%)",
      "midpoint": 2047,
      "prediction": "Aggregate 50% chance unaided machines outperform humans in every possible task (HLMI) by 2047; 10% by 2027.",
      "date": "2023-10",
      "type": "Expert survey",
      "source": {
        "title": "Grace et al.: Thousands of AI Authors on the Future of AI (arXiv)",
        "url": "https://arxiv.org/abs/2401.02843"
      }
    },
    {
      "person": "ESPAI 2023 (AI Impacts)",
      "org": "2,778 researchers at six top AI venues",
      "predicted": "2116 (50%)",
      "midpoint": 2116,
      "prediction": "Aggregate 50% chance all human occupations are fully automatable (FAOL) by 2116.",
      "date": "2023-10",
      "type": "Expert survey",
      "source": {
        "title": "Grace et al.: Thousands of AI Authors on the Future of AI (arXiv)",
        "url": "https://arxiv.org/abs/2401.02843"
      }
    },
    {
      "person": "ESPAI 2024 (AI Impacts)",
      "org": "1,580 researchers at six top AI venues",
      "predicted": "2042 (50%)",
      "midpoint": 2042,
      "prediction": "Mean aggregate forecast reaches a 50% chance of HLMI in 2042 (18 years after the survey) and 10% in 2027.",
      "date": "2024-12",
      "type": "Expert survey",
      "source": {
        "title": "AI Impacts: Advanced AI according to 1,580 researchers (ESPAI 2024, PDF)",
        "url": "http://aiimpacts.org/wp-content/uploads/2026/09/ESPAI2024.pdf"
      }
    },
    {
      "person": "ESPAI 2024 (AI Impacts)",
      "org": "1,580 researchers at six top AI venues",
      "predicted": "2098 (50%)",
      "midpoint": 2098,
      "prediction": "Mean aggregate forecast reaches a 50% chance of FAOL in 2098, 17 years earlier than the recomputed 2023 figure.",
      "date": "2024-12",
      "type": "Expert survey",
      "source": {
        "title": "AI Impacts: Advanced AI according to 1,580 researchers (ESPAI 2024, PDF)",
        "url": "http://aiimpacts.org/wp-content/uploads/2026/09/ESPAI2024.pdf"
      }
    },
    {
      "person": "Metaculus community",
      "org": "Forecasting platform, Q5121 (general AI)",
      "predicted": "2045 (median)",
      "midpoint": 2045,
      "prediction": "Community median date for the first general AI system to be devised, tested and publicly announced.",
      "date": "2020-12-27",
      "type": "Forecasting platform",
      "source": {
        "title": "Metaculus: Date of Artificial General Intelligence (Q5121)",
        "url": "https://www.metaculus.com/questions/5121/date-of-artificial-general-intelligence/"
      }
    },
    {
      "person": "Metaculus community",
      "org": "Forecasting platform, Q5121 (general AI)",
      "predicted": "2056 (median)",
      "midpoint": 2056,
      "prediction": "Community median for the first general AI system (about 112 forecasters).",
      "date": "2022-02-24",
      "type": "Forecasting platform",
      "source": {
        "title": "Metaculus: Date of Artificial General Intelligence (Q5121)",
        "url": "https://www.metaculus.com/questions/5121/date-of-artificial-general-intelligence/"
      }
    },
    {
      "person": "Metaculus community",
      "org": "Forecasting platform, Q5121 (general AI)",
      "predicted": "2038 (median)",
      "midpoint": 2038,
      "prediction": "Community median for the first general AI system after ChatGPT's release (about 346 forecasters).",
      "date": "2022-12-24",
      "type": "Forecasting platform",
      "source": {
        "title": "Metaculus: Date of Artificial General Intelligence (Q5121)",
        "url": "https://www.metaculus.com/questions/5121/date-of-artificial-general-intelligence/"
      }
    },
    {
      "person": "Metaculus community",
      "org": "Forecasting platform, Q5121 (general AI)",
      "predicted": "2031 (median)",
      "midpoint": 2031,
      "prediction": "Community median for the first general AI system (about 1,105 forecasters).",
      "date": "2023-12-21",
      "type": "Forecasting platform",
      "source": {
        "title": "Metaculus: Date of Artificial General Intelligence (Q5121)",
        "url": "https://www.metaculus.com/questions/5121/date-of-artificial-general-intelligence/"
      }
    },
    {
      "person": "Metaculus community",
      "org": "Forecasting platform, Q5121 (general AI)",
      "predicted": "2033 (median)",
      "midpoint": 2033,
      "prediction": "Community median for the first general AI system (about 1,664 forecasters).",
      "date": "2025-05-28",
      "type": "Forecasting platform",
      "source": {
        "title": "Metaculus: Date of Artificial General Intelligence (Q5121)",
        "url": "https://www.metaculus.com/questions/5121/date-of-artificial-general-intelligence/"
      }
    },
    {
      "person": "Metaculus community",
      "org": "Forecasting platform, Q5121 (general AI)",
      "predicted": "Sept 24, 2030 (median; IQR Mar 2028 to Aug 2036)",
      "midpoint": 2030,
      "prediction": "Community median for the first general AI system, as displayed on October 29, 2026 (about 1,840 forecasters).",
      "date": "2026-10-29",
      "type": "Forecasting platform",
      "source": {
        "title": "Metaculus: Date of Artificial General Intelligence (Q5121)",
        "url": "https://www.metaculus.com/questions/5121/date-of-artificial-general-intelligence/"
      }
    },
    {
      "person": "Metaculus community",
      "org": "Forecasting platform, Q3479 (weakly general AI)",
      "predicted": "2042 (median)",
      "midpoint": 2042,
      "prediction": "Community median date for the first weakly general AI system to be publicly known.",
      "date": "2022-02-22",
      "type": "Forecasting platform",
      "source": {
        "title": "Metaculus: Date weakly general AI is publicly known (Q3479)",
        "url": "https://www.metaculus.com/questions/3479/date-weakly-general-ai-is-publicly-known/"
      }
    },
    {
      "person": "Metaculus community",
      "org": "Forecasting platform, Q3479 (weakly general AI)",
      "predicted": "Nov 2027 (median)",
      "midpoint": 2027,
      "prediction": "Community median for the first weakly general AI system.",
      "date": "2022-12-27",
      "type": "Forecasting platform",
      "source": {
        "title": "Metaculus: Date weakly general AI is publicly known (Q3479)",
        "url": "https://www.metaculus.com/questions/3479/date-weakly-general-ai-is-publicly-known/"
      }
    },
    {
      "person": "Metaculus community",
      "org": "Forecasting platform, Q3479 (weakly general AI)",
      "predicted": "Aug 29, 2027 (median; IQR Sept 2026 to Oct 2029)",
      "midpoint": 2027,
      "prediction": "Community median for the first weakly general AI system, as displayed on October 29, 2026.",
      "date": "2026-10-29",
      "type": "Forecasting platform",
      "source": {
        "title": "Metaculus: Date weakly general AI is publicly known (Q3479)",
        "url": "https://www.metaculus.com/questions/3479/date-weakly-general-ai-is-publicly-known/"
      }
    },
    {
      "person": "Epoch AI Direct Approach model",
      "org": "Epoch AI",
      "predicted": "2029 (median; 10% 2025, 90% 2084)",
      "midpoint": 2029,
      "prediction": "Default-parameter output of the interactive model for transformative AI (TAI); 57% chance by 2030.",
      "date": "2023-05-31",
      "type": "Forecasting model",
      "source": {
        "title": "Epoch AI: Direct Approach interactive model",
        "url": "https://epoch.ai/publications/direct-approach-interactive-model"
      }
    }
  ]
}