A social constructivist perspective on long-term AI policy
I think the case for addressing the long-term consequences of AI systems holds even if AGI is unlikely to arise.
The future of AI development will be shaped by social, economic and political factors, and I’m not convinced that AGI will be desirable in the future or that AI is necessarily progressing toward AGI. However, (1) AI already has large positive and negative effects on society, and (2) I think it’s very likely that society’s AI capabilities will improve over time, amplifying these effects and creating new benefits and risks in the future.
A social constructivist perspective on long-term AI policy
I think the case for addressing the long-term consequences of AI systems holds even if AGI is unlikely to arise.
The future of AI development will be shaped by social, economic and political factors, and I’m not convinced that AGI will be desirable in the future or that AI is necessarily progressing toward AGI. However, (1) AI already has large positive and negative effects on society, and (2) I think it’s very likely that society’s AI capabilities will improve over time, amplifying these effects and creating new benefits and risks in the future.