In Silicon Valley, there are hopes that AI models will one day serve as effective therapists, providing mental health care without the requirements human counselors face, such as advanced degrees and continuous personal maintenance. However, recent reports indicate that the practical application of AI in therapeutic settings is not matching this optimistic vision.
A recent article detailed instances where therapists have been using ChatGPT during sessions without informing their clients. In one case, a therapist inadvertently revealed their screen during a virtual appointment, allowing the client to observe their private thoughts being generated by ChatGPT in real-time. This raises questions about the implications of using AI tools in therapy.
While the potential for AI to offer therapeutic benefits exists, the unmonitored and covert integration of unverified AI models into therapeutic practices is concerning. The risks associated with such practices also underscore the need for guidelines and ethical considerations in mental health care. A conversation between senior AI reporter James O’Donnell and journalist Laurie Clarke further examined this situation and its ramifications.
This story is part of “The Algorithm,” a weekly newsletter that covers developments in AI.
In related discussions on technological advancements, organizations are evaluating the transformative impact of emerging technologies. A LinkedIn Live event scheduled for September 10 will focus on various breakthroughs, including quantum computing, humanoid robotics, and climate technology, exploring both the innovations and challenges they present. Participants are invited to register for the event to engage in these discussions.
Source: https://www.technologyreview.com/2025/09/09/1123447/the-download-meet-our-ai-innovators-and-what-happens-when-therapists-use-ai-covertly/