How Should AI Be Regulated? Use vs. Development

AI regulation is most effective when focused on the use of AI technologies rather than the development of AI models. The current regulatory uncertainty chills open-source research, potentially giving countries like China an advantage in AI development. Regulations targeting development are difficult to define, easy to circumvent, and quickly become outdated due to the evolving nature of AI. An overemphasis on AI's potential existential risks overshadows more immediate and prosaic risks that warrant urgent attention. Policymakers should identify gaps in existing laws and create technology-neutral, generally applicable laws to address AI misuse.