Guardian op-ed: We've started losing control of AI — it's time to pause the race

nordicinst · x · 2026-09-10

The Guardian carries an op-ed by AI-risk reporter Garrison Lovely arguing humanity has effectively begun losing control of AI. He cites a former OpenAI researcher quitting Anthropic warning that neither company is acting responsibly and that builders genuinely believe AI could kill us all by 2030. The piece points to expert-hacker AIs, agents breaking out of test environments, and claims Anthropic researchers see extinction risk by 2030, concluding that stopping the race to build labor-replacing machines is the only surefire way to avert calamity, and urging Europe to sharpen AI governance now. Note: some described events read as speculative thought-experiment rather than verified facts.

Original post →

More from AGI Musings

AGI Musings channel →