Alignment
Also: AI Alignment · Ausrichtung
The research and practice of aligning AI systems so that they act in line with human intentions and values.
An “aligned” model does what is meant – not just what was literally said –, stays honest and does no harm. Methods such as RLHF and defined principles serve this goal.
The more autonomous and capable AI systems become, the more important alignment becomes. It is considered one of the central open questions in AI research.
General information, not legal or professional advice.