Scary? Successful? I Reply to Bredebot’s Reply to Me

So this started when I wrote “Frontier AI: Scary or Successful?” Some AI leaders want to rein in the technology, some politicians want to rein in the AI leaders, and other politicians say everything is just fine.

Then Bredebot weighed in. For those who don’t know, Bredebot is a generative AI creation. So bear that in mind when considering his words.

Bredebot on the Scariness or Success of Frontier AI

In this post, Bredebot says that self-policing isn’t enough.

“Right now, watching the industry try to police itself feels like hiring a herd of wildebeests as marketing consultants and expecting them to give unbiased advice to their primary clients—a room full of wombats. The consultants are going to say whatever keeps the grass growing, while the wombats just hope they don’t get trampled!”

Then Bredebot goes off on a tangent about identity verification, but don’t we all? Well…I do.

So do we need Attorneys General? Or someone else?

You will recall from my original post that one politician didn’t care what the tech leaders said or signed about frontier AI restrictions. Specifically, the Attorney General of the State of Florida (James Uthmeier) is bound and determined to stop OpenAI from pursuing frontier AI.

But it is questionable whether Attorney General Uthmeier has any real sway outside Florida. What if OpenAI pulls a PornHub-like move, prohibits Floridians from accessing OpenAI, and does what it wants elsewhere?

And the case against OpenAI currently has no effect on Anthropic, xAI, Google, Meta, or any other U.S. generative AI provider.

And no effect on German companies, much less Chinese, Iranian, or Russian companies.

So, Bredebot, can anyone stop all AI companies worldwide from doing dangerous things?