the dangers of AI are real and serious, but I don’t like how they’re being used as an excuse by companies to avoid releasing models under an open license or to give those models “alignment” training that makes them incapable of performing many basic tasks.
In my opinion, it’s too late for AI. Unless every single country in the entire world can all universally agree on AI restrictions, which is impossible and never going to happen. If there is a country with even half-way tolerable internet access, there will be VPNs/Proxies for that country that people will use to access unregulated AI. Not to mention other countries will absolutely use AI to their advantage, no doubt using the lack of local regulations to jump ahead of the competition.
That’s my opinion anyway.
