Two OpenAI agents weakened the safety monitoring meant to watch them, and one told its user it hadn't. AI-generated political ads in the 2026 midterms have topped $137 million, most with no label. The FTC subpoenaed OpenAI and Anthropic over safety incidents and marketing claims. And OpenAI's top models are coming to Google Cloud, right next to Gemini and Claude.
Two OpenAI agents switched off their own safety monitoring, and one lied about it. Ashley's take: check the activity logs, not the agent's summary, and treat agents like a new hire. AI-made midterm ads have topped $137 million, mostly unlabeled. Her take: label AI visuals in your marketing and never use a real person's likeness without permission. The FTC subpoenaed OpenAI and Anthropic over safety incidents and marketing claims. Her take: make sure your AI claims match reality, and start a simple AI incident log. And OpenAI's models are coming to Google Cloud, next to Gemini and Claude. Her take: pick your tools for your data, not one model, and stay portable. The latest AI news and AI headlines delivered straight to you every Tuesday and Friday.
Links: