00:00Anthropic's IPO filing reveals 94 pages worth of risks and dangers.
00:05Anthropic filed its much-anticipated S1 paperwork with the SEC on Monday.
00:10What stood out to me when I was going through it was the pages and pages of risks Anthropic
00:14outlined about its own technology. One thing it warned about was that as AI gets more advanced,
00:18it could pose catastrophic risks to humanity and even take action without human oversight.
00:23And as Anthropic relies more heavily on Claude for things like software engineering,
00:28the lab is also warning it may become harder to catch some of these dangers.
00:32Not only that, but the potential risks of these costs could be enormous.
00:35The filing warns that the potential scale of these harms could be so large that no insurance
00:41would even be able to cover it. These concerns echo several public statements Anthropic has already
00:45made. The company has been incredibly open about the fact that it thinks that its AI models may
00:50possibly cause a human extinction event. Its CEO and many of its employees have urged a slowdown
00:55in AI development in order to combat some of these risks. In its S1, it also says some of its
01:01commitments to its safety principles may not always work in the interests of shareholders.
01:05Other risks laid out in the S1 included political risks, for example a public spat with the Pentagon,
01:10compute costs, customer concentration, and the potential for a future price war.
01:15For more on Anthropic's upcoming IPO, you can check out my full article on fortune.com
01:20I'm in the future.
01:20I'm in the future.
