← Home
The Diary of a CEO with Steven Bartlett · July 13, 2026 · 2h 0m

OpenAI Whistleblower FINALLY Speaks: “AI Has A 70% Chance Of Going Horribly Wrong!“

Daniel Kokotajlo

Former OpenAI researcher Daniel Kokotajlo walked away from a $2 million non-disparagement agreement to speak out about the risks of artificial intelligence. He explains why he believes there's a 70% chance of human extinction from AI, why superintelligence could arrive before the end of this decade, and why almost every job is at risk. He also outlines the one plan he thinks could still lead to a safe and abundant future.

This summary was generated from show notes and public descriptions, not from a full transcript review. Details may contain inaccuracies.

Novel

•
The Plan A for Safe Superintelligence
Kokotajlo outlines a specific strategy in 'AI 2040: Plan A' to navigate superintelligence safely.
ClaudeChatGPTGemini

Highlights

•
Why He Left OpenAI00:13:04
Kokotajlo left OpenAI because of what he saw during ChatGPT's launch and his concerns about the direction of AI safety.
ClaudeChatGPTGemini
•
70% Chance of Human Extinction00:40:55
Kokotajlo estimates a 70% probability that AI leads to human extinction.
ClaudeChatGPTGemini
•
Superintelligence Before 203000:19:10
AI companies believe superintelligence could arrive sooner than publicly stated, possibly before the end of this decade.
ClaudeChatGPTGemini
•
Almost Every Job Could Be Automated00:47:38
AI could fully automate nearly all human jobs.
ClaudeChatGPTGemini

Misc

✧Kokotajlo walked away from $2 million rather than sign a non-disparagement agreement about his time at OpenAI.
Was this useful?