Coming Soon!
In my last blog there were warnings about using AI chatbots for therapy. Yet there is reason to hope that people will be able to protect themselves if they know that they are engaging with AI-generated responses. The self-generated protection is based on our affinity for empathy that comes from our fellow humans. We like it best from other pe0ple.
The 2025 Nov./Dec issue of the Monitor on Psychology reported on 9 studies presented in Nature Human Behaviour that included a total of 6,282 online subjects. In these studies identical empathic responses were labeled as coming from either AI or another person. Findings showed that the participants rated the responses from humans as more empathic and supportive and as evoking more positive and fewer negative reactions from them than those identified as coming from AI. When the participants were told that AI had contributed to the human empathic responses, their ratings of empathy and support fell significantly.These results were consistent across a variety of AI chatbots.
These studies strongly support our preference for real human engagement over responses from nonhumans. Human support is much more impactful than that generated from AI. These studies suggest that we are less likely to rely on AI for support and therapy than might be supposed.
Reference: In Brief (2025). Human empathy preferred over AI. Monitor on Psychology, 56(8), p. 17.