Social RoboticsOpenAI and MIT Media Lab study emotional use of ChatGPT
Heavy, emotionally expressive use was rare but linked to more loneliness and dependence.
OpenAI and the MIT Media Lab published early findings from an analysis of real ChatGPT conversations and a controlled trial on affective use and wellbeing. Most conversations contained no emotional content, but a small group of heavy voice-mode users were far more likely to treat the chatbot as a friend. Longer daily use was associated with worse outcomes, and people with a stronger tendency towards attachment fared worse, though the researchers stress these are correlations and the work was not peer reviewed. The study is one of the first large looks at how conversational AI, the same technology now built into social robots, affects loneliness and emotional reliance.
Source: MIT Media Lab
This archive entry summarises reporting published elsewhere; follow the source link for the full story. The summary was drafted with AI from that reporting.