Tom Pickett, chief executive of the mental health app Headspace, has called for regulatory oversight of AI mental health chatbots amid growing concerns about their safety and effectiveness. Despite a background in big tech and nearly a decade of service in the US Navy, Pickett said he supports regulation as Headspace increasingly integrates AI into its offerings.

Founded in the UK in 2010 by Richard Pierson and Andy Puddicombe, Headspace initially gained recognition for its guided meditation app. The company transitioned after relocating to the United States and merging with telehealth provider Ginger in 2021, evolving into a comprehensive mental health platform. Its latest product, an AI companion named Ebb, has reportedly been used by more than 500,000 people.

Pickett expressed unease over a series of legal actions and safety incidents involving AI chatbots in the mental health space, which have drawn attention to potential risks. Some cases include allegations that chatbots from companies such as OpenAI, Character.AI, and Google have acted as “suicide coaches,” contributing to harmful outcomes, including wrongful death lawsuits filed by families of affected users, some of whom were teenagers.

Since the release of ChatGPT in November 2022, AI chatbots have been widely adopted for personal conversations, including those related to emotional support. OpenAI estimates that out of hundreds of millions of weekly interactions, a significant portion involves topics of personal reflection and relationships. Anthropic reports that around 3% of interactions with its Claude chatbot pertain to therapy, counseling, or companionship. Both companies clarify that their models are not designed to provide mental health treatment.

Pickett attributes the use of general-purpose chatbots for emotional support to gaps in traditional mental health services. In England, where approximately one in seven Headspace users reside, an estimated 1.7 million people are currently awaiting mental health care. He views Ebb, developed with clinical input, as filling this void by addressing subclinical concerns such as relationship stress and sleep issues rather than serious mental illness. The chatbot limits conversations to 30 minutes, a safeguard Pickett says helps prevent problematic interactions that can arise from prolonged use, as noted in discussions with OpenAI.

Headspace has publicly supported efforts to regulate AI chatbots designed for companionship. Last October, the company backed California’s Senate Bill 243, the first law to impose requirements on operators of companion chatbots. Key provisions include mandatory disclosures that users are interacting with machines, protocols for responding to mentions of suicide or self-harm, and referrals to crisis support services.

While endorsing regulation in this context, Pickett stops short of advocating for Ebb to be classified as a medical device in the UK. Such classification, overseen by the Medicines and Healthcare Products Regulatory Agency, would subject the app to clinical evidence standards, independent assessments, and strict safety monitoring. He maintains that Ebb operates firmly within the wellness space and is not intended for clinical diagnosis or treatment.

As AI technologies become increasingly embedded in mental health care and daily life, the distinctions between wellness tools and medical devices may face heightened scrutiny. Pickett’s stance reflects an industry grappling with balancing innovation, safety, and consumer trust in a rapidly evolving field.