Big claims about AI, abundance, automation, and the future — tracked against reality over time.
Every AI era comes with bold forecasts: AGI by 2027, abundance by 2035, the singularity by 2045, mass automation, existential risk, radical longevity, robot labor, and post-scarcity economies.
This tracker collects the most interesting predictions and checks how they’re aging.
Leopold Aschenbrenner
AGI-level systems become plausible around 2027, followed by rapid superintelligence pressures.
Target
2027
Likelihood
40%
Capability: 45% · Deployment: 25%
Alexander Wissner-Gross & Peter Diamandis
AI unlocks compounding breakthroughs across scarcity, health, energy, and education.
Target
2035
Likelihood
30%
Capability: 55% · Deployment: 25%
Ray Kurzweil
Human-level AI arrives by 2029 and the technological singularity follows around 2045.
Target
2029
Likelihood
45%
Capability: 50% · Deployment: 25%
Kokotajlo, Alexander, Lifland, Larsen, et al.
Transformative AI progress becomes acute by 2027 under a detailed scenario timeline.
Target
2027
Likelihood
TBD
Dario Amodei
Powerful AI could compress a century of progress in several domains into roughly a decade.
Target
2035
Likelihood
TBD
Sam Altman
AI drives broad abundance and a more gradual singularity as intelligence becomes cheap.
Target
2035
Likelihood
TBD
Demis Hassabis
AGI may arrive within years rather than decades, with major scientific impact along the way.
Target
2030
Likelihood
TBD
Ajeya Cotra
Biological anchor methods imply material probability of transformative AI this century, with significant probability before mid-century.
Target
2050
Likelihood
TBD
Holden Karnofsky
This century could plausibly be the most important in history because of transformative AI.
Target
2100
Likelihood
TBD
Epoch AI
AI progress can be tracked through compute, data, algorithms, and benchmark trends.
Target
No specific date
Likelihood
TBD
Metaculus
Community forecasts provide question-specific probabilities for AI capabilities, impacts, and dates.
Target
No specific date
Likelihood
TBD
Katja Grace / AI Impacts
AI expert surveys estimate when high-level machine intelligence and related milestones may arrive.
Target
No specific date
Likelihood
TBD
Nick Bostrom
Once machine intelligence reaches human level, a rapid transition to superintelligence could create extreme control problems.
Target
No specific date
Likelihood
TBD
Eliezer Yudkowsky & Nate Soares
Building sufficiently advanced unaligned AI would be catastrophic, regardless of which actor builds it.
Target
No specific date
Likelihood
TBD
Stuart Russell
Advanced AI must be designed around uncertainty about human preferences to remain controllable.
Target
No specific date
Likelihood
TBD
Rodney Brooks
Many public AI and robotics forecasts are too aggressive and should be checked against dated scorecards.
Target
No specific date
Likelihood
TBD
Gary Marcus
Current scaling approaches are unlikely to produce robust AGI on the aggressive timelines often claimed.
Target
No specific date
Likelihood
TBD
Tony Seba
Technology cost curves drive rapid disruption across energy, transport, food, and related economic systems.
Target
2035
Likelihood
TBD
Martin Ford
AI and robotics will increasingly automate cognitive and physical work, pressuring wages and employment structures.
Target
No specific date
Likelihood
TBD
Vernor Vinge
The technological singularity may occur within roughly 30 years of the original 1993 essay.
Target
2023
Likelihood
TBD
