AI safety research

AI safety research aims to ensure AI systems behave as intended and avoid harmful outcomes. Alignment, interpretability, and robustness are key areas.

• AI safety ensures systems behave as intended.
• Alignment and robustness are key research areas.