Alignment

The effort to make AI systems behave in line with human intentions and values.

Why it matters

Alignment is the core of the AI safety debate: making powerful systems do what we actually want.