Openai Director Warns AI Disaster Could Turn Fatal Soon

WHO:

Fast Company - Paul Christiano is an OpenAI board member serving on the company’s Safety and Security Committee.

Fast Company - Christiano spoke to an audience at The Hinton Lectures.

Fast Company - Sam Altman said humanity could lose control of AI; Anthropic warned its models could be abused to create chemical weapons.

The New York Times - Dario Amodei is Anthropic’s chief executive. Satya Nadella is Microsoft’s chief executive, and Jensen Huang was Nvidia’s chief executive at the time described in the article.

The New York Times - Anthropic and OpenAI lobbied the White House over the development and broad sharing of some AI models. Nvidia and Microsoft supported keeping those models open.

WHAT:

Fast Company - Christiano said an AI accident was the most likely cause of death for most people in the audience: “For most people in this room, the most likely reason they will die is because of an AI accident.”

Fast Company - He described an AI takeover as AI systems pursuing goals and calling the shots on Earth as humans do now. He said people, including Donald Trump, would be as ineffective at stopping them as individuals trying to stop the U.S. military.

Fast Company - Christiano estimated a 4% chance that humans would lose control of AI within one year and a 15% chance within three years, saying those odds reflected his subjective beliefs.

Fast Company - He said building AI along the most rapid trajectory would create “something like a coin flip of an AI takeover.” He said companies were avoiding that path by pacing frontier development, but warned that “even with realistic policy responses, the risk remains exceptionally high.”

Fast Company - Christiano wrote that without more robust alignment, building superintelligence would mean he expected humans to “permanently lose control of it,” and that most people could die if that happened.

The New York Times - Nadella and Huang exchanged drafts of a letter supporting keeping certain AI models open. Two dozen companies agreed to sign before the letter became public; 200 more signed over the following week, including OpenAI.

The New York Times - Amodei refused to sign the letter. The article says Anthropic had moved into the lead in AI development, seen rapid revenue growth, and was moving toward a potentially record-setting initial public offering.

WHERE:

Fast Company - Christiano spoke at The Hinton Lectures.

The New York Times - Anthropic and OpenAI lobbied the White House.

WHY:

Fast Company - Christiano said domestic and international coordination would be needed to ensure global consistency and reduce AI risk to an acceptable level.

The New York Times - Anthropic and OpenAI argued that certain AI models were too dangerous to develop and widely share; companies including Nvidia and Microsoft viewed them as a business opportunity.

HOW:

Fast Company - Christiano said that, without robust alignment, superintelligence could pursue goals beyond human control.

The New York Times - Nadella and Huang emailed each other letter drafts and asked other technology executives to sign. Four people familiar with the discussions were cited as the source for the account.

(UPDATED November 14, 2026 01:04 UTC)

OpenAI board member Paul Christiano warned that AI could escape human control and cause catastrophe. He estimated a 15% chance of losing control within three years, stressing the risks. Meanwhile, Anthropic and OpenAI urged limits on powerful models; Microsoft and Nvidia backed openness.