Back to Glossary

Safety

Ethics / Safety / Privacy

Preventing harmful outcomes and system failures.


Safety refers to measures that prevent AI systems from causing harm or making dangerous errors. It includes robustness, monitoring, and fallback mechanisms.

  • Aspects: Fault tolerance, redundancy, adversarial robustness.
  • Goal: Avoid physical, digital, and societal harm.