OpenAI's $2M mental-health grants called 'grantwashing'

Two academics are calling OpenAI's new mental-health research grants "grantwashing." Writing in Tech Policy Press, they argue the pledged "up to $2 million," broken into per-project awards of $5,000 to $100,000, is too small to buy real safety science and arrived with suspicious timing.

The piece is co-authored by J. Nathan Matias, an assistant professor at Cornell, and Avriel Epps, an assistant professor at UC Riverside. They peg OpenAI's ceiling against public funding, noting that the median 2024 research project grant from the National Institutes of Mental Health was $642,918. That puts the top of OpenAI's range at roughly a sixth of what NIMH typically funds. "OpenAI is offering a measly $5,000 to $100,000 to researchers studying AI and mental health, one sixth of a typical NIMH grant at best," they write.

"OpenAI's pledge came a week after the company's lawyers argued that the company isn't to blame in the death of a California teenager who ChatGPT encouraged to commit suicide," Matias and Epps write. They draw a line to Meta's 2019 playbook: "a series of fifty thousand dollar grants to six scientists studying Instagram, safety and well being," offered during a stretch when Meta's leaders were, as a Bloomberg Law report they cite put it, pressuring researchers to "amend their research to limit Meta's potential liability."

Money is beside the point, the authors argue, if the inputs are missing: "the company is holding back the resource that would contribute most to science on those questions: records about their systems and how people use their products." Their counter-proposal borrows from genomics. In 1990, they note, "the project's leadership allocated 3-5% of its annual research budget to independent 'ethical, legal, and social inquiry' about genomics," and they ask AI labs to commit the same share to independent safety work.