AI may assist, but a human reviews anything involving customers, money, children, claims, private data, legal matters, health, hiring, or policy. Signal #34

TopCat AI Signal Report #34- Watch ListThe UN warned that AI is outpacing oversight and urged global rules, especially around child safety and AI systems that mimic friendship or manipulate young users. (Reuters⁠)The EU AI Act transparency rules take effect in August 2026, including requirements around AI interaction, deepfakes, and certain AI-generated public-interest content. (Digital Strategy⁠)The U.S. is still moving toward voluntary standards for advanced model release, keeping frontier AI inside a governance-and-review lane.