Tech Giants Urged to Share Insights on AI Uses to Enable Grounded Debate and Regulation
-
The latest AI models like GPT-4 are more general-purpose, making it hard to predict how they will be used and regulate effectively.
-
Even in regulated contexts like employment, many possible uses of generative AI models exist, but without information on actual use cases, it's difficult to evaluate risks.
-
Government mandates could require reporting on AI use in highly regulated industries, but will miss individual use cases outside of those contexts.
-
Tech companies building AI have the most comprehensive view of how their models are being used in practice, so they should share information to enable grounded debate on risks.
-
If tech companies want effective AI regulation, they need to voluntarily provide regulators information to foster understanding of general-purpose AI and real regulatory needs.