
I’m not against AI regulation at all, but I do think that any national AI regulation would risk preempting state laws because of the supremacy clause. That’s part of the reason we can’t pass a national privacy law (unless we make specific carve-outs for California and Illinois’ privacy laws). When each state has its own laws you significantly increase the work companies have to do to ensure compliance. A national framework is a lot easier
Although in the end, the state by state approach for privacy laws kind of worked itself out, because most states followed California’s model or Virginia and Colorado’s model. The lone exception is Maryland, which is wayyy different from most other laws. So companies only really have to check compliance with 3 models. You could argue that states might go this way with AI regulation too
If companies have to do more work, they’ll make even more of an effort to kill the bill or water it down, and pro-business politicians will help. And the supremacy clause absolutely means states can’t be stricter, unless specific provisions are made to defer when states like Illinois have stricter laws. In the past, we tried to handle all of these problems, and the bill became so watered down that civil rights and privacy advocates withdrew their support for the bill, and we ended up passing…
There’s a whole other issue because we haven’t actually adopted privacy as a basic constitutional right, while other countries have. It sounds wild, but we unfortunately have court precedent that says some uses of data are considered speech. Congress is understandably hesitant to make laws that could immediately be struck down, so lawmakers have to be super careful about how they discuss this and how it’s written