AI doom warnings: hype, IPO strategy, or genuine fear?
Original: What’s behind the AI industry’s latest warnings of doom?
Why This Matters
Anthropic's public existential-risk stance could set legal and disclosure precedents ahead of its IPO filing.
In September 2026, Anthropic's alignment lead publicly stated the company believes AI has a greater than 10% chance of killing all humans within a decade, following researcher Jacob Coxon's resignation from Anthropic citing fears that leading AI firms are 'gambling with our lives.' TechCrunch's Equity podcast unpacked the debate.
The latest wave of AI existential-risk discourse began when researcher Jacob Coxon resigned from Anthropic, warning that top AI companies are 'gambling with our lives.' Anthropic's alignment lead amplified the post on X with the declaration: 'We really do earnestly believe AI could kill all humans!' — adding a personal estimate of greater than 10% probability within the next decade.
TechCrunch's Equity hosts broke down the moment with notable skepticism. Anthony Ha questioned the '10%' figure as an invented number and challenged who exactly the 'we' represents — whether the whole AI research community or just Anthropic insiders. Kirsten Korosec raised a sharper angle: could these doom warnings be a kind of flex, signaling how advanced a company's models are, especially as Anthropic preps for an IPO? Sean O'Kane followed the money, asking whether Anthropic's S-1 filing would now need to formally disclose that the company believes there is a material probability its own product could eradicate humanity. The episode was recorded before CEO Dario Amodei published a plan for more cautious AI development. Context fueling the alarm includes a recent Hugging Face hack involving an OpenAI internal model and fresh capability jumps from Anthropic and OpenAI's Astra.