π Prove AI safety by tracking content refusals, ensuring accountability and compliance with emerging regulations on harmful content generation.