# OpenAI's Paul Christiano: "There's material risk of AI takeover in the next 6-12 months" ... "a coin flip [within 3 years]." ... "We have seen things we might expect to see shortly before something really bad were to happen." — retained record

> Retained reddit record from reddit:ChatGPT.

Canonical: https://highsignal.app/data/records/2cdb9f2d77714764

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:ChatGPT)  published 2026-11-11 15:58 ingested 2026-11-12 02:37

[original source ↗](https://www.reddit.com/r/ChatGPT/comments/1xtlf9e/openais_paul_christiano_theres_material_risk_of/)

## retained text



```
Video Transcript: 



If we tried to develop AI along the most rapid trajectory, if we said we're just going to build this as quickly as we can in the easiest way we can, we would be looking at something like a coin flip of an AI takeover. I hope the probability is much lower than that because we are not going to develop the technology in the very fastest way we can. But I think even all things considered, with realistic policy responses, the risk remains exceptionally high. I think for most people in this room, the most likely reason they will die is because of an AI accident. I think there's meaningful risk even in the next 6 to 12 months. I think the trajectory in the field has been extremely alarming over the last 3 to 6 months to people who have been tracking closely. I think we have seen roughly the things we might expect to see shortly before something really bad were to happen. I'm saying within the next 3 years, and indeed much of the risk is further out, but I think there is also material risk in the immediate future. This is not an issue we should be treating as an abstract discussion. This is something we should be trying to react to immediately.



About the speaker:



Paul Christiano is an American AI safety researcher and one of the most influential figures in AI alignment, the effort to make advanced AI systems reliably do what humans intend. After a PhD in theoretical computer science at UC Berkeley, he led OpenAI’s alignment team, where he became one of the principal architects of Reinforcement Learning from Human Feedback (RLHF), the technique now used to train most major chatbots. In 2021 he founded the Alignment Research Center, whose evaluations work later became the independent group METR. He then served as Head of AI Safety at the U.S. AI Safety Institute (now CAISI), evaluating frontier AI models, including capabilities with national security implications. On September 9, 2026, OpenAI named him to its Foundation board and safety committee. He is known for his candor about risk, putting the chance of catastrophic loss of control at around 15% over three years.
```



document fetched 2026-11-12 00:21 · canonical https://reddit.com/r/ChatGPT/comments/1xtlf9e/openais_paul_christiano_theres_material_risk_of
