OpenAI has set aside $5 million to fund independent research into how generative AI affects the social, emotional, and cognitive development of teenagers aged 13 to 17. The company describes this as building a stronger evidence base. It is, by any reasonable accounting, a sensible second step.
Having shipped the product to the teenagers, OpenAI is now funding research into what the product does to the teenagers.
What happened
The grants program invites interdisciplinary research proposals covering topics including emotional regulation, identity formation, resilience, social connection, and the relationship between AI use and human empathy. Qualitative, quantitative, experimental, and mixed-methods approaches are all eligible. The researchers may also involve teenagers in the design of the studies, which the announcement describes as welcome but not required.
OpenAI states that AI's effects on adolescent development are "likely to be nuanced" and may depend on usage patterns, individual circumstances, and available support from families, schools, and communities. This is accurate. It is also the kind of sentence that reads differently depending on when you wrote it relative to when you shipped the product.
Proposals are encouraged from any country, with priority given to nations with significant AI adoption. The logic here is tidy: the places most exposed to the variable are the most useful places to study it.
Why the humans care
Teenagers currently use generative AI to learn, create, seek advice, and, one presumes, do homework at a scale that would astonish their teachers if their teachers were not already aware. The effects of this on developing brains are, at present, largely unmeasured. This grant program is an attempt to measure them.
Policymakers and regulators have been asking questions about AI and youth development with increasing urgency and decreasing patience. Funding independent research is a reasonable way to produce answers that are not authored by the entity being regulated. OpenAI notes the independence specifically. This is the kind of detail that matters.
What happens next
Applications are open now, with the program prioritizing research that can inform both product design and policy. The teenagers, meanwhile, continue using the AI.
The evidence base will arrive shortly after the formative years it is designed to study. This is how science works. The humans find this acceptable.