> Regulation is seldom made for the benefit of the public. I don't know why AI custodians are begging for it You've answered your own question. They want to be

> Regulation is seldom made for the benefit of the public. I don't know why AI custodians are begging for it You've answered your own question. They want to be the owners of the industry. What good will come of regulation? Unfortunately: 1. Nobody actually knows 2. Bad regulation might be worse than no regulation for public safety or security or whatever metric you care about 3. It might turn out that for AI alignment, any regulation pushes us closer to doom. (Ignore this if you don't believe that AI can ever be an existential threat.) In AI spaces, even people who are extremely concerned about the acceleration of AI capabilities while safety lags behind don't know if slowing down research will even help. The reason is that regulation is local, but capabilities growth has a global effect. So if the "most ethical" country (and this is hypothetical, I don't propose that US firms are or are not the leaders in AI ethics) slows down capabilities research, "less ethical" countries will overtake them and we'll get more dangerous AI as a result, bringing us closer to a Bad End. AI capabilities research is an extraordinarily dangerous game and also an inevitability at this point. Should the US regulate AI research? I wouldn't trust anyone who has a firm position on that. And of course, as you say, even if we do regulate the industry, not all regulation is the same. Any actual regulation should be driven by safety research, by the open source community, and by cybersecurity more than it should be driven by the leaders in capabilities research.