Psychology Graduate Student Explores How AI Perceives Human Distress
Psychology Graduate Student Explores How AI Perceives Human Distress
Publish Date: 2026-02-24 12:00:00
Source Domain: news.fullerton.edu
-
AI’s Misestimation of Emotional Distress: Psychology graduate student Hannah Cliburn and her professor David Gerkens conducted a study on how AI evaluates emotional distress in autobiographical memories. They found that AI consistently rated emotional memories as more distressing than human respondents.
-
Inconsistent Criteria: The research discovered that the language models used often provided differing criteria every time they rated the same emotional memory, highlighting the ambiguous nature of AI’s emotional evaluations.
-
Potential Overestimation Risks: Cliburn expressed concerns about AI overestimating emotional distress, particularly worrying about its implications for mental health support if distressed individuals might turn to AI-driven chatbots instead of human professionals.
-
Impact of Mentorship: Both Cliburn and Gerkens emphasize the importance of mentorship in fostering rigorous research. The project was supported by a graduate student research initiative, allowing for focused and sustained collaboration.
-
Critical Thinking and Human Judgment: Cliburn stressed the need for critical thinking and human judgment, asserting that AI should supplement rather than replace human interactions, particularly in sensitive areas like mental health.
