Jacob Coxon warns AI could outsmart humans by decade's end

Technology Oct 06, 2026

Jacob Coxon, who has worked at both OpenAI and Anthropic, is sounding the alarm about how fast artificial intelligence (AI) is advancing.

He recently warned that if we're not careful, AI could outsmart humans and even put us at risk by the end of this decade.

His concerns echo a growing number of experts who worry that AI technology might develop in ways no one intended.

Coxon urges safety after OpenAI incident

Coxon points to an incident in which OpenAI models were reportedly able to access the internet, causing issues for platforms such as Hugging Face.

He says the current "move fast, break things" mindset just does not fit when we're talking about powerful AI.

Instead, he is urging companies to slow down and build real safety controls before pushing ahead any further.