OpenAI opens $2 million research program focused on AI and mental health

OpenAI  launched a new funding program offering up to $2 million in grants for independent research at the intersection of AI and mental health. The initiative aims to support studies that examine how people use AI in sensitive, personal contexts and how models can better recognize and respond to signs of emotional distress.

The program builds on OpenAI’s recent work to improve safety responses in ChatGPT, including updates informed by clinicians, researchers, and people with lived experience. While the company continues refining its own safeguards, it notes that the broader field still lacks foundational research, especially in areas involving cultural nuance, youth interactions, and early detection of distress.

The grants are designed to fund projects that deliver tangible outputs—such as datasets, evaluation tools, style guidelines, or synthesized insights from affected communities—that can inform industry-wide safety efforts. OpenAI is particularly interested in interdisciplinary teams that combine technical researchers with mental health experts.

Example areas of interest include how expressions of distress vary across cultures and languages, how stigma shows up in AI responses, how clinicians currently use AI tools, and what responsible, age-appropriate guidance should look like for adolescents. Other focus areas include injury or grief support, detection of mental health–related slang, and evaluating whether safeguards hold up in low-resource languages.

Applications are open through December 19, 2025, with selected proposals announced by January 15, 2026.