In an interview for Harvard Magazine, he discusses the legal and cybersecurity ramifications of large (and powerful) AI models, and the ongoing debate over legal liabilities for emerging technologies.