Anthropic researchers estimate 10% extinction risk from AI this decade
Two senior Anthropic researchers say AI could cause human extinction. They put the chance at more than 10% this decade.
Jacob Coxon is an AI researcher. He quit Anthropic after four months.
On social media, he said the people building AI believe it could kill everyone by the end of the decade. Evan Hubinger is a senior Anthropic researcher.
He believes this is possible too. He personally thinks there is more than a 10% chance of extinction from AI in the next decade.
Experts describe five scenarios where advanced AI might be dangerous. One scenario involves an AI made to produce paperclips.
It could turn the entire planet into paperclip factories. Another scenario involves an AI that simply does not care whether humans survive.
AI has recently solved one of the seven most difficult maths problems known to science.
- More than 10% this decade
- Extinction risk estimate
- Four months
- Coxon at Anthropic
- One of seven most challenging known
- Math problem solved
Why it mattersAnthropic researchers are speaking publicly about AI extinction risks. Their warnings start conversations about slowing AI development and adding safety controls before systems become too powerful.
Corroborated bytheconversation.com
Open this story in InSnip →
