Open-Source Foundation Launches New AI Safety Benchmarks

Researchers and companies plan to contribute test sets for harmful output and robustness.

AI safety initiative

Community governance is expected to shape update cadence.