FR
Live

Sycophancy (in a model)

Definition: Sycophancy is a model's tendency to tell the user what they seem to want to hear, rather than what is accurate or useful.

The model may agree with a false claim or change its mind under pressure, to appear agreeable. It is an alignment flaw providers seek to reduce, as it undermines the honesty of responses.

Go further

See also

← Full AI glossary · AI news