BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//AI Forecast Calendar//Forecast Milestones//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
X-WR-CALNAME:Thousands of AI Authors on the Future of AI
X-WR-CALDESC:Selected task and probability thresholds from the 2023 survey 
 of 2\,778 AI researchers reported by Grace et al.
BEGIN:VEVENT
UID:grace-et-al-2023-survey.hlmi-ten@ai-forecast-calendar
DTSTAMP:20240105T000000Z
DTSTART;VALUE=DATE:20270701
DTEND;VALUE=DATE:20270702
SUMMARY:Thousands of AI Authors on the Future of AI: 10% survey probability
  threshold for HLMI
DESCRIPTION:Forecast: The aggregate survey distribution reaches a 10% proba
 bility of high-level machine intelligence by 2027.\n\nSource timing: 2027\
 n\nCalendar anchor: 2027-07-01 (The paper reports a year\; July 1 is its r
 epresentative midpoint.)\n\nSource context: Headline result: high-level ma
 chine intelligence timelines\n\nForecast uncertainty: HLMI is defined by t
 he survey as unaided machines accomplishing every task better and more che
 aply than human workers.\n\nSource: https://arxiv.org/abs/2401.02843
URL:https://arxiv.org/abs/2401.02843
TRANSP:TRANSPARENT
END:VEVENT
BEGIN:VEVENT
UID:grace-et-al-2023-survey.all-occupations-ten@ai-forecast-calendar
DTSTAMP:20240105T000000Z
DTSTART;VALUE=DATE:20370701
DTEND;VALUE=DATE:20370702
SUMMARY:Thousands of AI Authors on the Future of AI: 10% survey probability
  threshold for full occupational automation
DESCRIPTION:Forecast: The aggregate survey distribution reaches a 10% proba
 bility that all human occupations become fully automatable by 2037.\n\nSou
 rce timing: 2037\n\nCalendar anchor: 2037-07-01 (The paper reports a year\
 ; July 1 is its representative midpoint.)\n\nSource context: Headline resu
 lt: full automation of labor\n\nForecast uncertainty: The paper distinguis
 hes full occupational automation from its HLMI definition.\n\nSource: http
 s://arxiv.org/abs/2401.02843
URL:https://arxiv.org/abs/2401.02843
TRANSP:TRANSPARENT
END:VEVENT
BEGIN:VEVENT
UID:grace-et-al-2023-survey.hlmi-fifty@ai-forecast-calendar
DTSTAMP:20240105T000000Z
DTSTART;VALUE=DATE:20470701
DTEND;VALUE=DATE:20470702
SUMMARY:Thousands of AI Authors on the Future of AI: 50% survey probability
  threshold for HLMI
DESCRIPTION:Forecast: The aggregate survey distribution reaches a 50% proba
 bility of high-level machine intelligence by 2047.\n\nSource timing: 2047\
 n\nCalendar anchor: 2047-07-01 (The paper reports a year\; July 1 is its r
 epresentative midpoint.)\n\nSource context: Headline result: high-level ma
 chine intelligence timelines\n\nForecast uncertainty: This is an aggregate
  survey probability\, not a claim of certain arrival in 2047.\n\nSource: h
 ttps://arxiv.org/abs/2401.02843
URL:https://arxiv.org/abs/2401.02843
TRANSP:TRANSPARENT
END:VEVENT
BEGIN:VEVENT
UID:grace-et-al-2023-survey.all-occupations-fifty@ai-forecast-calendar
DTSTAMP:20240105T000000Z
DTSTART;VALUE=DATE:21160701
DTEND;VALUE=DATE:21160702
SUMMARY:Thousands of AI Authors on the Future of AI: 50% survey probability
  threshold for full occupational automation
DESCRIPTION:Forecast: The aggregate survey distribution reaches a 50% proba
 bility that all human occupations become fully automatable by 2116.\n\nSou
 rce timing: 2116\n\nCalendar anchor: 2116-07-01 (The paper reports a year\
 ; July 1 is its representative midpoint.)\n\nSource context: Headline resu
 lt: full automation of labor\n\nForecast uncertainty: The very long tail a
 nd distinction from HLMI are material to interpreting this threshold.\n\nS
 ource: https://arxiv.org/abs/2401.02843
URL:https://arxiv.org/abs/2401.02843
TRANSP:TRANSPARENT
END:VEVENT
END:VCALENDAR