Self-regulation of the AI industry can be effective. But ensuring safety, security, and public trust will require collaboration with major stakeholders, including the U.S. government.