Earlier today (4/9), the Center for AI Policy released and promoted a bill titled the "Responsible Advanced Artificial Intelligence Act." It calls for many limits on AI models.

The bill was featured on Politico Pro. Here's a paragraph:

The draft bill is so far without any congressional sponsors. But Green-Lowe said that CAIP — which launched last year and registered its first lobbyist in December — has convened around 80 meetings on Capitol Hill to discuss what it’s calling the Responsible Advanced AI Act. The organization has also held four briefings for congressional staff on the topic, Green-Lowe said.

The promotion of the bill to outlets without any supporters portrays AI governance and alignment in a negative light. This felt like it was trying to shift the Overton window, but it may end up backfiring and making it recede. Although the bill had a very unusual rollout since it didn't have supporters, this was a high-integrity action.

https://twitter.com/neil_chilson/status/1777695468656505153

Would be curious what other people thought.

New to LessWrong?

New Comment