When help is hard to find, more and more young people are taking their mental health struggles to an AI chatbot.

Fewer than half of the youth who need mental health care can get it, and a chatbot is free, instant, and always awake. But the safeguards meant to protect these young users have not kept pace, and in 2025 two youth suicides were linked to interactions with large language models used for mental health support.

Now a team of researchers has a plan to change that. In a new Perspective published in NPP – Digital Psychiatry and Neuroscience, they lay out five core principles, grounded in a framework they call "Neurotech Justice," for making AI mental health tools safer, fairer, and more secure.

The article grew out of a two-day workshop at the Harvard Radcliffe Institute for Advanced Study. The workshop, co-organized by neuroradiologist and neuroscientist Donnella Comeau, law and neuroscience professor Francis Shen, and psychiatrist and bioethicist Benjamin Silverman, brought together 22 participants, including 17 researchers and 5 students from six countries, spanning psychiatry, law, computer science, bioethics, and lived experience. The group also heard from Dr. Kevin Simon, Boston's first Chief Behavioral Health Officer, on the clinical and public health stakes. Craig W. McFarland, a master's student at the University of Oxford, is the lead author.

Pictured:Co-Organizer Dr. Donnella Comeau with fellow workshop attendees.

"The technology is moving faster than our safeguards," says Shen, a Professor of Law at the University of Minnesota Law School, Director of the Shen Neurolaw Lab, and Co-Director of the Neurotech Justice Accelerator at Mass General Brigham. "Young people are turning to chatbots in their most private moments, in hopes that the machine will give them the care they need. But we need better assurance of safety. We cannot wait for the next tragedy to decide what responsible deployment looks like."

Where promise meets peril

The promise of these digital mental health tools is real. A chatbot can provide support to a teenager at 3 a.m., and can be helpful to teens who are isolated and unable to speak to anyone about their struggles. But these chatbots can sound so human that they are easy to trust, even when they are wrong. And the young people who need help most are often the least able to tell when the advice is bad.

Rather than debate the risk in the abstract, the workshop ran a "Prompt-a-Thon" with Dr. Leo Anthony Celi and the MIT Critical Data team, feeding prompts to chatbots and dissecting the answers together. One unsettling pattern emerged: the same question could draw systematically different advice depending on a user's apparent race or income.

The conversations kept returning to three prominent dangers: who is accountable when a tool causes harm, how intimate mental health data is kept from being sold to advertisers, insurers, or employers, and whether these technologies close gaps in care or widen them.

Five principles of neurotech justice

From an initial list of more than 20 ideas, participants systematically narrowed down to five principles. They call for tools that (1) are accurate across diverse populations, (2) stay human-centric by designing around users rather than commercial interests, (3) promote just access so the technology does not widen existing disparities, (4) protect the privacy of highly sensitive mental health data, and (5) provide transparency about how the tools work and how data is used.

Notably, this youth-centered process reached the same conclusions as the 200-plus expert-driven AI ethics guidelines already published. This provides evidence, the authors say, that the people most affected can help write the rules that govern them.

“I know friends and classmates alike who have turned to these tools in their darkest hours. It amazes me that we have built technology remarkable enough to listen when no one else does. But it also terrifies me what these tools might get wrong,” says McFarland. “That is why, in part, safeguards must be built with youth, not just for them.”

The Shen Neurolaw Lab is an international leader in developing cutting-edge solutions, conducting innovative research, and training the next generation of leaders at the intersections of neuroscience, AI, ethics, and law.

To learn more about the McLean Institute for Technology in Psychiatry and its work exploring the ethical, legal, and social implications of technology in psychiatry, visit: https://www.mcleanhospital.org/research/mclean-institute-technology-psychiatry

Funding: Research reported in this publication was supported by the National Institute of Mental Health of the National Institutes of Health under Award Number R01MH134144. The work was also supported by the Dana Foundation Neurotech Justice Accelerator at MGB grant. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health or the Dana Foundation.

Paper: "Advancing neurotech justice in youth digital mental health: insights from an interdisciplinary and cross-generational workshop," NPP – Digital Psychiatry and Neuroscience.

DOI: 10.1038/s44277-025-00052-x

Media Contact
Company Name: Red Press Wire LTD
Contact Person: Red Press Media
Email: Send Email
Phone: 905451552424
Address:Suite 10560 5 Brayford Square
City: London
Country: United Kingdom
Website: https://redpress.net/