Can a Chatbot Really Support Your Mental Health? Experts Say It’s Closer Than You Think

As global demand for mental health support continues to soar, researchers in the US are working hard to bring credibility to AI-generated therapy tools. With traditional therapists in short supply, tech innovators and psychologists are looking toward AI to fill the gap, ethically and effectively.

Dartmouth College is one of the key players in this mission. Their team, led by Nick Jacobson, has developed Therabot, an AI-powered application aimed at delivering trustworthy psychotherapy. The goal isn’t just to create another mental health app, but one that separates itself from flashy, untested tools by rooting itself in clinical science and safety protocols.

Therabot – More Than Just Another Mental Health App

Therabot addresses the severe shortage of trained therapists. Jacobson, an assistant professor at Dartmouth, points out that even a tenfold increase in therapists would still fall short of meeting demand. Therabot’s approach involves carefully crafted simulations of patient-caregiver interactions to ensure the AI doesn’t rely on generic or harmful content.

Clinical trials are already underway, with early results showing promise for people experiencing anxiety, depression, and eating disorders. Another trial is set to compare Therabot’s results directly with those of conventional therapies.

The Push for Ethics, Safety, and Transparency

The developers have spent over six years refining Therabot, with safety and effectiveness placed above speed or profits. Michael Heinz, co-leader of the project and a psychiatrist, emphasizes that rushing a mental health tool to market without rigorous testing could cause more harm than good.

They’re even considering launching a nonprofit wing to keep therapy accessible for those who can’t afford in-person treatment, adding another layer of social responsibility to the mission.

Concerns Around AI in Therapy – What Experts Are Saying

Vaile Wright of the American Psychological Association envisions a future with ethically developed AI tools supporting mental health—but she stresses they must be rooted in science and co-created by experts. She highlights the danger of apps that simply say what users want to hear, especially for vulnerable young people who may not realize they’re being manipulated.

Darlene King from the APA adds that while AI holds potential, we still lack enough data to understand its long-term benefits and risks fully.

Regulation? Not Quite There Yet

Despite growing interest, AI therapy tools still sit in a regulatory grey zone. The FDA doesn’t certify these apps as medical devices, though it may approve them for marketing after a review process. This gap leaves users with little formal protection if something goes wrong.

Always Available, But Not Always Ideal

Herbert Bay, CEO of Earkick, is also exploring this space with his app, Panda, which alerts users to emotional crises. He contrasts his approach with the case of Character.AI, which was linked to a tragic incident involving a teenager.

Still, Bay believes AI therapists are best suited for daily mental health support rather than handling severe crises. One user, Darren, found comfort using ChatGPT during a period of trauma—though it wasn’t built for therapy, it still helped him feel heard.

Conclusion

AI in mental health isn’t about replacing therapists—it’s about widening access, especially for those in need. However, success will depend on building systems with caution, transparency, and scientific grounding. As AI continues its march into healthcare, one thing is clear: trust will be the ultimate test.