Eli Lifland

Quick Assessment

Key Links

Overview

Eli LiflandPersonEli LiflandBiographical profile of Eli Lifland, a top-ranked forecaster and AI safety researcher who co-authored the AI 2027 scenario forecast and co-founded the AI Futures Project. The page documents his for...Quality: 58/100 is a forecaster and AI safety researcher who ranks #1 on the RAND Forecasting Initiative all-time leaderboard. He co-leads the SamotsvetyOrganizationSamotsvetyElite forecasting group Samotsvety dominated INFER competitions 2020-2022 with relative Brier scores twice as good as competitors, providing influential probabilistic forecasts including 28% TAI by...Quality: 61/100 forecasting team, which placed first in the CSET-Foretell/INFER competition in 2020, 2021, and 2022.¹ His work focuses on AGI timelineConceptAGI TimelineComprehensive synthesis of AGI timeline forecasts showing dramatic acceleration: expert median dropped from 2061 (2018) to 2047 (2023), Metaculus from 50 years to 5 years since 2020, with current p...Quality: 59/100 forecasting, scenario planning, and AI safety.

Lifland co-founded the AI Futures ProjectOrganizationAI Futures ProjectAI Futures Project is a nonprofit co-founded in 2024 by Daniel Kokotajlo, Eli Lifland, and Thomas Larsen that produces detailed AI capability forecasts, most notably the AI 2027 scenario depicting ...Quality: 50/100 alongside Daniel Kokotajlo and Thomas Larsen, and co-authored AI 2027, a detailed scenario forecast exploring potential AGI developmentE604Comprehensive synthesis of AGI timeline forecasts showing dramatic compression: Metaculus aggregates predict 25% probability by 2027 and 50% by 2031 (down from 50-year median in 2020), with industr...Quality: 52/100 trajectories.²³ The project, with contributions from Scott Alexander and Romeo Dean, provides a concrete scenario for how superhuman AI capabilities might emerge, including geopolitical tensions, technical breakthroughs, and alignment challenges.

Lifland also co-founded and advises Sage, an organization building interactive AI explainers and forecasting tools, and serves as a guest fund manager at the Long Term Future Fund.⁴ He previously worked on ElicitOrganizationElicit (AI Research Tool)Elicit is an AI research assistant with 2M+ users that searches 138M papers and automates literature reviews, founded by AI alignment researchers from Ought and funded by Coefficient Giving (former...Quality: 63/100 at Ought and co-created TextAttack, a Python framework for adversarial attacks in natural language processing.⁵

AI Futures Project and AI 2027

Lifland is a co-founder and researcher at the AI Futures Project, a 501(c)(3) organization focused on AGI forecasting, scenario planning, and policy engagement.⁶ The organization was co-founded with Daniel Kokotajlo (Executive Director, former OpenAIOrganizationOpenAIComprehensive organizational profile of OpenAI documenting evolution from 2015 non-profit to Public Benefit Corporation, with detailed analysis of governance crisis, 2024-2025 ownership restructuri...Quality: 62/100 researcher) and Thomas Larsen (founder of the Center for AI Policy).⁷

The project's flagship output is AI 2027, a detailed scenario forecast released in April 2025 exploring how superintelligence might emerge.⁸ The scenario was co-authored with Scott Alexander (who primarily assisted with rewriting) and Romeo Dean (who contributed supplements on compute and security considerations).⁹

The AI 2027 forecast presents a concrete narrative of AI development including:

• Increasingly capable AI agents automating significant portions of AI research and development¹⁰
• Geopolitical tensions, particularly a US-China AI race, influencing safety decisions and deployment timelines¹¹
• Alignment challenges, including exploration of safer model series using chain-of-thought reasoning to address failures¹²
• Economic impacts, including widespread job displacement¹³

The project received significant attention and has been discussed in venues including Lawfare Media, ControlAIOrganizationControlAIControlAI is a UK-based advocacy organization that has achieved notable policy engagement success (briefing 150+ lawmakers, securing support from 100+ UK parliamentarians) while promoting direct in...Quality: 63/100, and a CEPR webinar.¹⁴¹⁵¹⁶

AI Futures Timelines Model

The AI Futures Project maintains a quantitative timelines model that generates probability distributions for key AGI milestones such as Automated Coder (AC) and superintelligence (ASI). The model incorporates benchmark tracking, compute availability, algorithmic progress, and other inputs to produce forecasts that team members then adjust based on their individual judgment.¹⁷

Lifland's personal AGI timeline estimates have shifted as new evidence has emerged. His median TED-AI (a general intelligence milestone) forecast has followed this trajectory:¹⁸

• 2021: ~2060
• July 2022: ~2050
• January 2024: ~2038
• Mid-2024: ~2035
• December 2024: ~2032
• April 2025: ~2031
• July 2025: ~2033
• January 2026: ~2035

The AI Futures Project has emphasized that the AI 2027 scenario was never intended as a confident prediction that AGI would arrive in 2027, and that all team members maintain high uncertainty about when AGI and ASI will be built.¹⁹ The December 2025 model update predicted 3-5 years longer timelines to full coding automation compared to the April 2025 AI 2027 forecast, attributed primarily to more conservative modeling of pre-automation AI R&D speedups and recognition of potential data bottlenecks.²⁰

Forecasting Track Record

Lifland ranks #1 on the RAND Forecasting Initiative (CSET-Foretell/INFER) all-time leaderboard.²¹ On GJOpen, his Brier score of 0.23 outperforms the median of 0.301 (ratio 0.76), and he secured 2nd place in the MetaculusOrganizationMetaculusMetaculus is a reputation-based forecasting platform with 1M+ predictions showing AGI probability at 25% by 2027 and 50% by 2031 (down from 50 years away in 2020). Analysis finds good short-term ca...Quality: 50/100 Economist 2021 tournament and 1st in the Salk Tournament as of September 2022.²²

As co-lead of the Samotsvety Forecasting team (approximately 15 forecasters), Lifland helped guide the team to first-place finishes in the INFER competition in 2020, 2021, and 2022.²³ In 2020, Samotsvety placed 1st with a relative score of -0.912 compared to -0.062 for 2nd place. In 2021, they achieved 1st with a relative score of -3.259 compared to -0.889 for 2nd place. Samotsvety holds positions 1, 2, 3, and 4 in INFER's all-time ranking, with some members achieving Superforecaster status.²⁴

The team has produced public forecasts on critical topics including AI existential risk and nuclear risk.²⁵

Sage and AI Digest

Lifland co-founded Sage, an organization focused on building interactive AI explainers and forecasting tools.²⁶ One of Sage's key projects is AI Digest, which received $550,000 from Coefficient GivingOrganizationCoefficient GivingCoefficient Giving (formerly Open Philanthropy) has directed \$4B+ in grants since 2014, including \$336M to AI safety (~60% of external funding). The organization spent ~\$50M on AI safety in 2024...Quality: 55/100 for its work, with an additional $550,000 for forecasting projects.²⁷ The organization aims to make AI developments more accessible to broader audiences through interactive tools and clear explanations.

Role in the AI Safety Community

Lifland is active in the AI safety and alignment communities, particularly through LessWrongOrganizationLessWrongLessWrong is a rationality-focused community blog founded in 2009 that has influenced AI safety discourse, receiving \$5M+ in funding and serving as the origin point for ~31% of EA survey responden...Quality: 44/100 and the Effective Altruism Forum. He serves as a mentor in the MATS Program (focusing on Strategy & Forecasting, Policy & Governance streams).²⁸ He has also been featured in the documentary "Making God," which explores AGI risks.²⁹

Lifland has taken the Giving What We Can Pledge, committing to donate 10% of his lifetime income to effective charities.³⁰

Criticisms and Controversies

Lifland's work, particularly the AI 2027 timelines model, has faced methodological criticism from community members. In a detailed critique posted to LessWrongOrganizationLessWrongLessWrong is a rationality-focused community blog founded in 2009 that has influenced AI safety discourse, receiving \$5M+ in funding and serving as the origin point for ~31% of EA survey responden...Quality: 44/100, the EA Forum, and Substack, forecaster "titotal" described the model's fundamental structure as "highly questionable," with little empirical validation and poor justification for parameters like superexponential time horizon growth curves.³¹ Titotal argued that models need strong conceptual and empirical justifications before influencing major decisions, characterizing AI 2027 as resembling a "shoddy toy model stapled to a sci-fi short story" disguised as rigorous research.³²

Critics have also raised concerns about philosophical overconfidence, warning that popularizing flawed models could lead people to make significant life decisions based on shaky forecasts.³³ Others counter that inaction on short timelines could be costlier if the forecasts prove accurate.³⁴

Lifland responded to these criticisms by acknowledging errors and reviewing titotal's critique for factual accuracy. He agreed to changes in the model write-up and paid $500 bounties to both titotal and another critic, Peter Johnson, for identifying issues.³⁵³⁶ The team released a detailed response explaining their reasoning more thoroughly, including their justification for the model's assumptions.³⁷

Other criticisms include:

• Lack of skeptic engagement: Some community members felt AI 2027 did not sufficiently address skeptical frameworks or justify its models against competing views³⁸
• Unverifiable predictions: Concerns that some predictions are difficult to validate empirically³⁹

Lifland has been forthright about forecast misses and has regularly updated his timelines as new evidence emerges.⁴⁰ No major personal controversies or ethical issues have been documented beyond these methodological debates.

Sources