Building AI That Respects Our Privacy

A personal commentary argues for integrating privacy into AI by shifting to model-per-user and on-device closed systems, increasing transparency through data watermarking, and enforcing data removal rights, while offering near-term user practices such as limiting shared data, understanding platform data use, recognizing AI’s limitations, and maintaining situational awareness against chatbot-driven scams until stronger legal protections exist.