Novoselva

This is really, really high.

It is relatively easy to create this self-fulfilling prophecy.

Way back when, in the AI dark ages of, oh, 2021, some OpenAI execs left to form what they called the Safer Alternative, Anthropic.

Owen Daniels is with Georgetown University's Center for Security and Emerging Technology.

To avoid things like AI that lies to achieve a goal or that builds bioweapons on command.

Core to Anthropic's safety effort had been a pledge called the Responsible Scaling Policy, says Sarah Myers-West of the AI Now Institute.

Anthropic says it was trying to stave off a race to a bottom with that policy.

But it admits it failed to persuade other companies to take its stricter safety approach.

The change it announced yesterday means they may keep building more advanced models, even if they're riskier.

Brent Thill, a tech industry analyst at Jefferies, says AI companies are getting more ambitious.

You need more advanced AI to run financial systems versus finding the best latte.

And as companies spend hundreds of billions of dollars to build those tools, Georgetown's Owen Daniels says they need to start making money, too.

While balancing that with the risks AI may pose.

An Anthropic spokesperson said they're now committing to more transparency to publish detailed reports about the risks and capabilities of its AI.

I'm Novoselva for Marketplace.