Glossary
AI alignment
The effort to make AI systems behave in accordance with intended human goals, values and constraints.
In the book
We distinguish the wider research and safety problem from getting useful assistance in an everyday exchange. Clarifying your purpose can improve a conversation; it does not settle the wider problem of designing and governing safe systems.
Introduced in chapter 3, “ResponsAIbility”.
Related terms
- Misalignment: A mismatch between an AI system’s behaviour or objectives and the human intentions or requirements it should serve.