Media & Interviews · Georgetown
Georgetown psychology colloquium examines chatbot empathy and sycophancy
A campus lecture explored why language models can appear empathetic while validating users without enough context or challenge.
What was discussed
The Hoya reports that Desmond Ong presented research on large language models that people may perceive as highly empathetic even when the systems follow repetitive validation patterns or fail to challenge harmful assumptions. The Oct. 2 Georgetown Psychology colloquium also examined the difference between chatbot responses and the contextual judgment of human relationships and licensed therapists.
Evidence boundary
This briefing summarizes The Hoya’s reporting. Hoyas.ai did not verify a complete recording or transcript of the colloquium and does not treat the article as clinical guidance.