Back
Clarifying How Our AI Timelines Forecasts Have Changed Since AI 2027
webblog.ai-futures.org·blog.ai-futures.org/p/clarifying-how-our-ai-timelines-for...
A January 2026 clarification post from the AI Futures Project addressing media misreporting on updated AGI timelines; relevant context for understanding AI 2027 scenario forecasting and the epistemic uncertainty the authors emphasize around transformative AI arrival dates.
Metadata
Importance: 52/100blog postcommentary
Summary
The AI Futures Project (Eli Lifland, Daniel Kokotajlo, Brendan Halstead) corrects media misrepresentations of their updated AI timelines forecasts, clarifying that they were never confident AGI would arrive in 2027 and explaining how their probabilistic estimates have evolved since publishing AI 2027 in April 2025. The post provides detailed breakdowns of past and current median/modal predictions for AGI, superintelligence, and related milestones, alongside a new AI Futures Model.
Key Points
- •The authors were never confident AGI would arrive in 2027; media reports mischaracterized modal predictions as confident point forecasts.
- •Major outlets (Guardian, WaPo, Independent) incorrectly compared old modal predictions to new median predictions, overstating apparent timeline shifts.
- •Daniel Kokotajlo's current all-things-considered distribution for transformative AI milestones is presented with updated uncertainty ranges.
- •The post tracks how Eli and Daniel's AGI median estimates have evolved over time, showing gradual rather than dramatic shifts.
- •The AI Futures Model (aifuturesmodel.com) now defaults to Daniel's parameters rather than Eli's to reduce public confusion.
Cited by 2 pages
| Page | Type | Quality |
|---|---|---|
| AI Timelines | Concept | 95.0 |
| Eli Lifland | Person | 58.0 |
Cached Content Preview
HTTP 200Fetched Apr 9, 202614 KB
Clarifying how our AI timelines forecasts have changed since AI 2027
AI Futures Project
Subscribe Sign in Clarifying how our AI timelines forecasts have changed since AI 2027
Correcting common misunderstandings
Eli Lifland , Daniel Kokotajlo , and Brendan Halstead Jan 27, 2026 59 10 8 Share Some recent news articles discuss updates to our AI timelines since AI 2027, most notably our new timelines and takeoff model, the AI Futures Model (see blog post announcement ). 1 While we’re glad to see broader discussion of the AI timelines, these articles make substantial errors in their reporting. Please don’t assume that their contents accurately represent things we’ve written or believe! This post aims to clarify our past and current views. 2
The articles in question include:
The Guardian: Leading AI expert delays timeline for its possible destruction of humanity
The Independent: AI ‘could be last technology humanity ever builds’, expert warns in ‘doom timeline’
Inc: AI Expert Predicted AI Would End Humanity in 2027—Now He’s Changing His Timeline
WaPo: The world has a few more years
Daily Mirror: AI expert reveals exactly how long is left until terrifying end of humanity
Our views at a high level
Important things that we believed in Apr 2025 when we published AI 2027, and still believe now:
AGI and superintelligence (ASI) will eventually be built and might be built soon, and thus we should be prepared for them to be built soon.
We are highly uncertain about when AGI and ASI will be built, we certainly cannot confidently predict a specific year.
How exactly have we changed our minds over the past 9 months? Here are the highlights.
Here is Daniel’s current all-things-considered distribution for TED-AI ( source ):
If you’d like to see a more complete table including more metrics as well as our model’s raw outputs, see below .
We’ve also made this graph of Daniel and Eli’s AGI medians over time, which goes further into the past:
See below for the data behind this graph.
Correcting common misunderstandings
Categorizing the misunderstandings/misrepresentations in articles covering our work:
Implying that we were confident an AI milestone (e.g. SC, AGI, or ASI) would happen in 2027 (Guardian, Inc, Daily Mirror) . We’ve done our best to make it clear that it has never been the case that we were confident AGI would arrive in 2027. For example, we emphasized our uncertainty several times in AI 2027 and, to make it even more clear, we’ve recently added a paragraph explaining this to the AI 2027 foreword.
Comparing our old modal prediction to our new model’s predi
... (truncated, 14 KB total)Resource ID:
6ce1c19011d01e71 | Stable ID: YjM1YzZmOG