A coalition of researchers calls on AI developers to transparently share safety methods, evaluations, and data with the global scientific community to enable independent scrutiny and advance AI safety research. Transparency should include safety-relevant details about frontier models while allowing exceptions for genuine security risks.
The Institute for Justice challenges Flock Safety's claims that its automatic license plate reader cameras reduce car theft, finding that crime trends in deploying jurisdictions were already improving before camera installation and questioning whether results hold up when data are weighted equally rather than favoring agencies with higher initial theft rates.