00:02The company behind one of the world's most powerful AI chatbots says some scientists have
00:07been using it for research that could help develop biological weapons, and Anthropic says it blocked
00:13their accounts. The company released a new report Thursday detailing how people have misused its
00:18clawed AI models. Anthropic identified five cases involving scientists doing work that had the
00:24potential to support biological weapons development. The company doesn't identify the
00:29scientists and says it can't say they intended to cause harm. Legitimate biological research can
00:35also produce information that could be used to make a pathogen more dangerous. Anthropic says it
00:40has since strengthened safeguards in its newer models. The report also details clawed being used
00:46for propaganda and surveillance, including by actors Anthropic says were aligned with governments in
00:51Iran and China. Anthropic says it's disclosing the cases because, quote, as models become increasingly
00:57capable, their risks will increase unless AI developers and society's defenders act to make
01:03them safer. The report comes days after an Anthropic researcher, Jacob Coxon, quit. Appearing on CNN,
01:10he says he's worried about what happens as these systems become more powerful. If you extrapolate
01:15into the future the level of capabilities of these AIs with the same independent volition they could
01:20cause extreme havoc. For example, hacking critical infrastructure, building extinction level
01:27bioweapons. There's a lot of ways that the AI could actuate itself in the world.
Comments