AI Glossary

AI agent safety

Also called agent safety, agentic AI safety

Keeping an AI agent that acts on real systems from causing harm: people keep control before high-stakes decisions, the agent's work stays visible to them, and the system guards data and resists misuse. At Orbyt Labs the term covers agents doing real work, not research into the values of the models themselves.

Source: Anthropic 2025 (opens in a new tab)

Read more: AI agent safety method

See also

  • False refusal

    A safety control refusing a request that is actually harmless, often because the request resembles an unsafe one or mentions a sensitive topic.

Cite this term.

CC BY 4.0, attribution required.

This term

Orbyt Labs, AI agent safety (AI Glossary), Nov 14, 2026. https://www.orbytlabs.ai/orbyt-collective/glossary/ai-agent-safety

The dataset (JSON)

Orbyt Labs Glossaries (Nov 14, 2026), https://www.orbytlabs.ai/glossary#dataset

Last updated .