The US Federal Trade Commission is investigating whether security incidents involving OpenAI and Anthropic models harmed consumers. Its demands reportedly contain more than 50 pages of questions on incidents, safeguards and marketing; targeted third parties may also be subpoenaed, potentially suppor