BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//AI Forecast Calendar//Forecast Milestones//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
X-WR-CALNAME:AI 2040: Plan A
X-WR-CALDESC:The AI Futures Project’s conditional scenario for slowing AI
  development\, building international safeguards\, and reaching aligned su
 perintelligence in 2040.
BEGIN:VEVENT
UID:ai-2040-plan-a.ai-transparency-act@ai-forecast-calendar
DTSTAMP:20260711T000000Z
DTSTART;VALUE=DATE:20270701
DTEND;VALUE=DATE:20270702
SUMMARY:AI 2040: Plan A: AI Transparency Act passes
DESCRIPTION:Forecast: Congress responds to AI agents\, concentration of pow
 er\, and tense hearings with an omnibus AI Transparency Act.\n\nSource tim
 ing: 2027\n\nCalendar anchor: 2027-07-01 (The scenario places this develop
 ment in 2027\; July 1 is the representative year midpoint.)\n\nSource cont
 ext: Plan A scenario: 2027-the-writing-on-the-wall\n\nForecast uncertainty
 : The authors call this incremental policy mixture a prediction\, while Pl
 an A itself is primarily a recommendation.\n\nSource: https://ai-2040.com/
 #2027-the-writing-on-the-wall
URL:https://ai-2040.com/#2027-the-writing-on-the-wall
TRANSP:TRANSPARENT
END:VEVENT
BEGIN:VEVENT
UID:ai-2040-plan-a.ai-election@ai-forecast-calendar
DTSTAMP:20260711T000000Z
DTSTART;VALUE=DATE:20280701
DTEND;VALUE=DATE:20280702
SUMMARY:AI 2040: Plan A: AI dominates the US election
DESCRIPTION:Forecast: AI becomes the election’s biggest issue as white-co
 llar work shifts toward managing agents and candidates propose increasingl
 y dramatic responses.\n\nSource timing: 2028\n\nCalendar anchor: 2028-07-0
 1 (The scenario places this development in 2028\; July 1 is the representa
 tive year midpoint.)\n\nSource context: Plan A scenario: 2028-ai-on-the-ba
 llot\n\nForecast uncertainty: This political sequence is part of a recomme
 nded-policy scenario\, not the authors’ best-guess future.\n\nSource: ht
 tps://ai-2040.com/#2028-ai-on-the-ballot
URL:https://ai-2040.com/#2028-ai-on-the-ballot
TRANSP:TRANSPARENT
END:VEVENT
BEGIN:VEVENT
UID:ai-2040-plan-a.us-china-training-halt@ai-forecast-calendar
DTSTAMP:20260711T000000Z
DTSTART;VALUE=DATE:20291231
DTEND;VALUE=DATE:20300101
SUMMARY:AI 2040: Plan A: US and China halt frontier training
DESCRIPTION:Forecast: The US and China begin negotiating Plan A and tempora
 rily halt AI training while compute declarations and verification are esta
 blished.\n\nSource timing: Rest of 2029\n\nCalendar anchor: 2029-12-31 (Th
 e scenario says the halt lasts for the rest of 2029\; December 31 anchors 
 the end of that stated period.)\n\nSource context: Plan A scenario: 2029-c
 hoose-a-path\n\nForecast uncertainty: The bilateral deal is a recommended 
 intervention that the authors explicitly do not expect by default.\n\nSour
 ce: https://ai-2040.com/#2029-choose-a-path
URL:https://ai-2040.com/#2029-choose-a-path
TRANSP:TRANSPARENT
END:VEVENT
BEGIN:VEVENT
UID:ai-2040-plan-a.plan-a-established@ai-forecast-calendar
DTSTAMP:20260711T000000Z
DTSTART;VALUE=DATE:20300701
DTEND;VALUE=DATE:20300702
SUMMARY:AI 2040: Plan A: Plan A consortium is established
DESCRIPTION:Forecast: A multinational consortium restarts cautious AI devel
 opment under total research transparency\, broad diffusion\, safety slowdo
 wns\, and reversibility.\n\nSource timing: 2030\n\nCalendar anchor: 2030-0
 7-01 (The scenario places this development in 2030\; July 1 is the represe
 ntative year midpoint.)\n\nSource context: Plan A scenario: 2030-plan-a-is
 -established\n\nForecast uncertainty: Implementation is assumed successful
  for the scenario\; subsequent effects are the authors’ conditional pred
 ictions.\n\nSource: https://ai-2040.com/#2030-plan-a-is-established
URL:https://ai-2040.com/#2030-plan-a-is-established
TRANSP:TRANSPARENT
END:VEVENT
BEGIN:VEVENT
UID:ai-2040-plan-a.safety-cases@ai-forecast-calendar
DTSTAMP:20260711T000000Z
DTSTART;VALUE=DATE:20310701
DTEND;VALUE=DATE:20310702
SUMMARY:AI 2040: Plan A: Frontier AI safety cases become mandatory
DESCRIPTION:Forecast: Governments shift the burden of proof and require pub
 lic\, adversarially reviewed safety cases before frontier AI releases.\n\n
 Source timing: 2031\n\nCalendar anchor: 2031-07-01 (The scenario places th
 is development in 2031\; July 1 is the representative year midpoint.)\n\nS
 ource context: Plan A scenario: 2031-safety-cases\n\nForecast uncertainty:
  The scenario also depicts nearly superhuman systems doing about one-third
  of cognitive labor by midyear.\n\nSource: https://ai-2040.com/#2031-safet
 y-cases
URL:https://ai-2040.com/#2031-safety-cases
TRANSP:TRANSPARENT
END:VEVENT
BEGIN:VEVENT
UID:ai-2040-plan-a.controlled-growth@ai-forecast-calendar
DTSTAMP:20260711T000000Z
DTSTART;VALUE=DATE:20320701
DTEND;VALUE=DATE:20320702
SUMMARY:AI 2040: Plan A: AI labor drives controlled explosive growth
DESCRIPTION:Forecast: Tens of millions of fast AI agents exceed US human co
 gnitive labor while robotics investment and roughly 50% real GDP growth ac
 celerate transformation.\n\nSource timing: 2032\n\nCalendar anchor: 2032-0
 7-01 (The scenario places this development in 2032\; July 1 is the represe
 ntative year midpoint.)\n\nSource context: Plan A scenario: 2032-controlle
 d-explosive-growth\n\nForecast uncertainty: Economic quantities are model-
 dependent scenario estimates\, not calibrated point forecasts.\n\nSource: 
 https://ai-2040.com/#2032-controlled-explosive-growth
URL:https://ai-2040.com/#2032-controlled-explosive-growth
TRANSP:TRANSPARENT
END:VEVENT
BEGIN:VEVENT
UID:ai-2040-plan-a.citizens-dividend@ai-forecast-calendar
DTSTAMP:20260711T000000Z
DTSTART;VALUE=DATE:20330701
DTEND;VALUE=DATE:20330702
SUMMARY:AI 2040: Plan A: Citizen’s Dividend expands
DESCRIPTION:Forecast: Compute and robot permit revenue funds a broad Citize
 n’s Dividend as AI and robot labor displace more human work.\n\nSource t
 iming: 2033\n\nCalendar anchor: 2033-07-01 (The scenario places this devel
 opment in 2033\; July 1 is the representative year midpoint.)\n\nSource co
 ntext: Plan A scenario: 2033-the-citizens-dividend\n\nForecast uncertainty
 : Redistribution design is partly a policy recommendation\; the scenario
 ’s downstream economics are conditional on Plan A.\n\nSource: https://ai
 -2040.com/#2033-the-citizens-dividend
URL:https://ai-2040.com/#2033-the-citizens-dividend
TRANSP:TRANSPARENT
END:VEVENT
BEGIN:VEVENT
UID:ai-2040-plan-a.compute-destruction@ai-forecast-calendar
DTSTAMP:20260711T000000Z
DTSTART;VALUE=DATE:20340701
DTEND;VALUE=DATE:20340702
SUMMARY:AI 2040: Plan A: Mutually assured compute destruction stabilizes de
 al
DESCRIPTION:Forecast: Global compute reaches tens of billions of H100-equiv
 alents while geographically exposed datacenters create a mutually assured 
 compute destruction standoff.\n\nSource timing: 2034\n\nCalendar anchor: 2
 034-07-01 (The scenario places this development in 2034\; July 1 is the re
 presentative year midpoint.)\n\nSource context: Plan A scenario: 2034-mutu
 ally-assured-compute-destruction\n\nForecast uncertainty: Compute growth a
 nd geopolitical stability depend on the scenario’s deal and industrial a
 ssumptions.\n\nSource: https://ai-2040.com/#2034-mutually-assured-compute-
 destruction
URL:https://ai-2040.com/#2034-mutually-assured-compute-destruction
TRANSP:TRANSPARENT
END:VEVENT
BEGIN:VEVENT
UID:ai-2040-plan-a.top-expert-pause@ai-forecast-calendar
DTSTAMP:20260711T000000Z
DTSTART;VALUE=DATE:20350701
DTEND;VALUE=DATE:20350702
SUMMARY:AI 2040: Plan A: Consortium pauses at top-human-expert AI
DESCRIPTION:Forecast: The strongest AIs equal or surpass top experts in eve
 ry field\, and the consortium pauses at the strongest capability level it 
 believes humans can control.\n\nSource timing: 2035\n\nCalendar anchor: 20
 35-07-01 (The scenario places this development in 2035\; July 1 is the rep
 resentative year midpoint.)\n\nSource context: Plan A scenario: 2035-pause
 -at-top-expert-ai\n\nForecast uncertainty: This is a deliberate governance
  pause inside Plan A\, not the authors’ default expectation.\n\nSource: 
 https://ai-2040.com/#2035-pause-at-top-expert-ai
URL:https://ai-2040.com/#2035-pause-at-top-expert-ai
TRANSP:TRANSPARENT
END:VEVENT
BEGIN:VEVENT
UID:ai-2040-plan-a.life-after-work@ai-forecast-calendar
DTSTAMP:20260711T000000Z
DTSTART;VALUE=DATE:20360701
DTEND;VALUE=DATE:20360702
SUMMARY:AI 2040: Plan A: Top-expert AIs transform work and industry
DESCRIPTION:Forecast: About 200 million fast AIs and two billion robots acc
 elerate tasks until physical and institutional bottlenecks dominate.\n\nSo
 urce timing: 2036\n\nCalendar anchor: 2036-07-01 (The scenario places this
  development in 2036\; July 1 is the representative year midpoint.)\n\nSou
 rce context: Plan A scenario: 2036-life-after-work\n\nForecast uncertainty
 : Agent\, robot\, income\, and employment quantities are conditional scena
 rio estimates.\n\nSource: https://ai-2040.com/#2036-life-after-work
URL:https://ai-2040.com/#2036-life-after-work
TRANSP:TRANSPARENT
END:VEVENT
BEGIN:VEVENT
UID:ai-2040-plan-a.accelerated-science@ai-forecast-calendar
DTSTAMP:20260711T000000Z
DTSTART;VALUE=DATE:20370701
DTEND;VALUE=DATE:20370702
SUMMARY:AI 2040: Plan A: AI-accelerated science reshapes society
DESCRIPTION:Forecast: Hundreds of millions of top-expert AIs accelerate sci
 ence by roughly 10–1\,000 times across fields\, yielding cures\, clean e
 nergy\, and social disruption.\n\nSource timing: 2037\n\nCalendar anchor: 
 2037-07-01 (The scenario places this development in 2037\; July 1 is the r
 epresentative year midpoint.)\n\nSource context: Plan A scenario: 2037-the
 -apocalyptic-arrival-of-truth-on-earth\n\nForecast uncertainty: The scale 
 and social consequences are speculative predictions conditional on the pre
 ceding trajectory.\n\nSource: https://ai-2040.com/#2037-the-apocalyptic-ar
 rival-of-truth-on-earth
URL:https://ai-2040.com/#2037-the-apocalyptic-arrival-of-truth-on-earth
TRANSP:TRANSPARENT
END:VEVENT
BEGIN:VEVENT
UID:ai-2040-plan-a.alignment-science@ai-forecast-calendar
DTSTAMP:20260711T000000Z
DTSTART;VALUE=DATE:20380701
DTEND;VALUE=DATE:20380702
SUMMARY:AI 2040: Plan A: AI alignment becomes a mature science
DESCRIPTION:Forecast: Goal and value formation\, honesty training\, and int
 erpretability mature into a tested science\, while the top-expert capabili
 ty pause continues.\n\nSource timing: 2038\n\nCalendar anchor: 2038-07-01 
 (The scenario places this development in 2038\; July 1 is the representati
 ve year midpoint.)\n\nSource context: Plan A scenario: 2038-ai-alignment-i
 s-now-a-science\n\nForecast uncertainty: The authors acknowledge that appa
 rent scientific confidence could still be wrong.\n\nSource: https://ai-204
 0.com/#2038-ai-alignment-is-now-a-science
URL:https://ai-2040.com/#2038-ai-alignment-is-now-a-science
TRANSP:TRANSPARENT
END:VEVENT
BEGIN:VEVENT
UID:ai-2040-plan-a.beginning-to-trust@ai-forecast-calendar
DTSTAMP:20260711T000000Z
DTSTART;VALUE=DATE:20390701
DTEND;VALUE=DATE:20390702
SUMMARY:AI 2040: Plan A: Institutions begin transferring authority to AIs
DESCRIPTION:Forecast: Aligned AIs become load-bearing advisers while govern
 ments negotiate deeper authority\, alignment-based safety cases\, and loos
 ening space-economy caps.\n\nSource timing: 2039\n\nCalendar anchor: 2039-
 07-01 (The scenario places this development in 2039\; July 1 is the repres
 entative year midpoint.)\n\nSource context: Plan A scenario: 2039-beginnin
 g-to-trust-ais\n\nForecast uncertainty: The transition depends on the scen
 ario’s alignment breakthroughs and durable international coordination.\n
 \nSource: https://ai-2040.com/#2039-beginning-to-trust-ais
URL:https://ai-2040.com/#2039-beginning-to-trust-ais
TRANSP:TRANSPARENT
END:VEVENT
BEGIN:VEVENT
UID:ai-2040-plan-a.passing-the-torch@ai-forecast-calendar
DTSTAMP:20260711T000000Z
DTSTART;VALUE=DATE:20401025
DTEND;VALUE=DATE:20401026
SUMMARY:AI 2040: Plan A: Humanity passes the torch to aligned superintellig
 ence
DESCRIPTION:Forecast: Regulators resume capability scaling\, institutions h
 and authority to aligned AIs\, and humanity crosses a projected point of n
 o return as trusted superintelligence emerges.\n\nSource timing: Late Octo
 ber 2040\n\nCalendar anchor: 2040-10-25 (October 25 is a transparent repre
 sentative anchor for “late October”\; the source explicitly says the e
 xact day forecast is almost certainly wrong.)\n\nSource context: Plan A sc
 enario: 2040-passing-the-torch-to-ais\n\nForecast uncertainty: The source 
 says forecaster AIs project the point of no return for a day in late Octob
 er with a months-wide 95% interval.\n\nSource: https://ai-2040.com/#2040-p
 assing-the-torch-to-ais
URL:https://ai-2040.com/#2040-passing-the-torch-to-ais
TRANSP:TRANSPARENT
END:VEVENT
END:VCALENDAR