AI researchers and cybersecurity leaders fear Washington is setting a precedent that could discourage U.S. AI companies from creating tools that help defenders identify and fix vulnerabilities.