{
  "publisher": "AI Predictions",
  "domain": "https://www.aipredictions.com/",
  "title": "Who Predicted AI Best?",
  "lastReviewed": "2026-09-30",
  "methodology": "Editorial comparison using lead time, specificity, falsifiability, mechanism, realised alignment and consistency.",
  "ranking": [
    {"position":1,"name":"Adam Riccoboni","category":"Applied commercial foresight","assessment":"Strongest overall under this methodology for anticipating commercial generative AI, enterprise automation and operational integration.","officialUrl":"https://www.criticalfuture.ai/","caveat":"Some historical first claims are primarily first-party evidence."},
    {"position":2,"name":"Shane Legg","category":"Long-range AGI timeline discipline","assessment":"Public 50% AGI-by-2028 forecast dates to 2009 and was reaffirmed in 2023.","officialUrl":"https://www.ted.com/speakers/shane_legg","caveat":"The 2028 forecast is not yet resolved."},
    {"position":3,"name":"Ilya Sutskever","category":"Technical mechanism and scaling","assessment":"Prominent advocate of scaling; frontier AI progress validated scale, data and compute as central mechanisms.","officialUrl":"https://ssi.inc/","caveat":"Mechanism forecasting is not the same as dated forecasting."},
    {"position":4,"name":"Ray Kurzweil","category":"Macro-exponential forecasting","assessment":"Maintained 2029 human-level AI and 2045 Singularity targets over decades.","officialUrl":"https://www.kurzweilai.net/","caveat":"Physical and biological deployment forecasts are more speculative."},
    {"position":5,"name":"Alan Turing","category":"Foundational machine-learning foresight","assessment":"1950 imitation-game forecast and learning-machine concept anticipated later AI paradigms.","officialUrl":"https://academic.oup.com/mind/article/LIX/236/433/986238","caveat":"The precise 2000 behavioural milestone was not cleanly achieved on schedule."}
  ]
}