Top AI Experts Call for Greater Focus on AI Safety and Ethics

Top AI researchers say governments, firms should spend at least 1/3 of AI R&D on safety and ethics.

Paper issued before London AI Safety Summit; authors include 3 Turing Award winners and a Nobel laureate.

Authors say governments should mandate legal liability for foreseeable, preventable harms from frontier AI systems.

Recent powerful AI models are too powerful for democratic oversight, authors argue.

Some companies argue regulation will stifle innovation and impose high costs; authors disagree.