OpenAI puts up $5M to study how AI affects teens

Teens today are not just using AI to cheat on their history essays. They are using chatbots to learn, brainstorm art projects, and vent about their daily anxieties. But nobody actually knows what growing up alongside an always-available, infinitely patient synthetic companion does to a developing human brain.

Now, OpenAI is putting some cash on the table to find out.

OpenAI committed $5 million in grants to support independent research on how generative AI impacts young people ages 13 to 17. The program offers individual grants of up to $1 million, funded directly by OpenAI Group PBC.

Probing the teenage mind

Applications are open through October 6, 2026, with decisions expected by November 13. The focus areas zero in on emotional well-being and social connection.

Specifically, researchers are encouraged to explore how AI affects resilience, self-esteem, empathy, and family relationships. OpenAI also wants hard data on whether current safeguards actually work across different cultures, languages, and socioeconomic groups.

There are a few catches worth noting. While the initiative welcomes interdisciplinary teams from psychology to data science, OpenAI explicitly stated that preference will go to projects studying ChatGPT, underrepresented populations, and work that can deliver actionable evidence for near-term policy and regulatory debates.

Recipients cannot take their time, either. Grantees must deliver an interim progress report in Q1 2027 and produce a working draft suitable for conference discussion within 12 months. Overhead costs are strictly capped at 10%.

Why it matters

For researchers and educators, this is a massive influx of capital into adolescent tech studies. High-dollar grants without traditional academic red tape are rare, making this an immediate opportunity for child-development labs wanting to test real-world AI usage.

For the AI industry, the regulatory clock is ticking loudly. Regulators worldwide are already circling social media companies over youth mental health. AI labs know they are next in line. Funding this research gives OpenAI a head start in building an evidence base before lawmakers force top-down restrictions.

We think this is a shrewd defensive play, though not without obvious trade-offs. By capping overhead at 10% and requiring rapid turnarounds, OpenAI is buying fast, targeted data that will directly inform its product design and policy lobbying. However, giving preference to ChatGPT-centric studies risks turning independent inquiry into product-specific feedback. If OpenAI truly wants bulletproof credibility, it will have to ensure critical, uncomfortable findings do not get quietly swept aside.