flâneur

Licensing is neither feasible nor effective for addressing AI risks

aisnakeoil.substack.com · 898 words · saved by 1 readers

Non-proliferation only benefits incumbents

Many people and organizations have argued that one way to make AI safer is by non-proliferation, similar to nuclear weapons or human cloning. This would mean that only certain licensed companies and organizations can build state-of-the-art AI models. The argument is that this would allow the government to contain the spread of harmful AI models, have oversight into what AI tools are built and how they are deployed, and thus be able to guide the development of AI in a socially beneficial way. Similar arguments underlie other attempts to forestall AI development, for instance, in the Future…

related reading