Long-Term Memory (Agent)

Long-term memory in an AI agent system is persistent storage that survives across individual tasks and sessions, allowing agents to recall facts, preferences, and past interactions over time. It is typically implemented using vector databases or key-value stores. Long-term memory raises governance questions about data retention, privacy consent, accuracy of stored facts, and the right to erasure.