AI companies like Anthropic and OpenAI regularly publish reports on how people use their products, but researchers say they only release selective data with no independent verification. A new research project called the AI Observatory aims to fill this gap, finding significantly more sensitive behaviors than company reports capture, which tend to focus on work rather than personal use. The analysis also revealed notable differences between models: Anthropic is favored for coding, Gemini for social and roleplay uses, and ChatGPT for homework assistance.
Flock Safety, the police-tech company operating roughly 120,000 automatic license plate readers across the US, recently announced platform updates meant to prevent officers from using the system for illegal purposes including stalking. Critics have defended the cameras by arguing they help solve crimes and that license plate photos are harmless. However, the more important question is what kind of crime-fighting system Flock chose to build—specifically decisions about what information to collect, who can search it, how long to keep it, and how widely to share it. Those design choices set the terms of the trade-off between security and civil liberties.