OpenAI AI and Teen Development Research Grant 2026: Up to $1 Million per Project
OpenAI is offering up to $1 million per project from a $5 million fund for independent research on generative AI and the development, wellbeing, and experiences of teens ages 13–17.
This captured cycle appears closed. Use this page for historical guidance unless the official source has reopened the program.
Captured cycle: This page is retained for historical guidance. Confirm whether the program has reopened before planning an application.
OpenAI AI and Teen Development Research Grant 2026: Up to $1 Million per Project
OpenAI’s 2026 AI and Teen Development Research Grant Program is a new call for independent research into how generative AI affects young people ages 13–17. Applications are open now and due October 6, 2026, at 11:59 p.m. Pacific Daylight Time. OpenAI Group PBC has committed a total of $5 million to the program and says individual awards may be as large as $1 million.
The call is international and deliberately broad in discipline and method. It invites work in areas including adolescent development, psychology, public health, human-computer interaction, communication, sociology, anthropology, data science, and AI safety. A strong proposal will connect a precise question about young people’s lives to an ethical, feasible study and to findings that can inform product safeguards, public policy, or future research. The applicant does not need to promise a favorable conclusion about AI: the program expressly values credible work regardless of whether its findings benefit technology providers.
Key details
| Detail | Information |
|---|---|
| Opportunity | OpenAI AI and Teen Development Research Grant Program |
| Cycle | 2026 call for research proposals |
| Funder and administrator | OpenAI Group PBC |
| Total program commitment | $5 million |
| Maximum individual grant | Up to $1 million |
| Application deadline | October 6, 2026, 11:59 p.m. PDT |
| Applicant age | 18 or older |
| Applicant profile | Research institution or organization affiliation and/or significant relevant experience |
| Geographic scope | Applicants from any country are encouraged; countries with significant AI adoption may be prioritized |
| Subject focus | Generative AI and teen development, centered on young people ages 13–17 |
| Indirect costs | No more than 10% of a project grant |
| Application language | English; research may be conducted in any language |
| Official announcement | OpenAI grant announcement and requirements |
| Submission portal | OpenAI application portal |
| Questions | [email protected] |
What the grant supports
The program is intended to build evidence about how teens use generative AI and how different patterns of use may relate to developmental outcomes. OpenAI frames the call around the complexity of teenage experiences: effects may vary with purpose, amount and context of use, age, individual circumstances, and the support provided by caregivers, schools, communities, and technology providers. The grant therefore has room for studies that measure benefits, risks, mixed effects, or differences between groups and settings.
Research may look at social and emotional development, with examples such as emotion regulation, resilience, identity, agency, self-esteem, social connection, friendship, family relationships, belonging, empathy, communication, and social skills. The call also names technical and practical questions: whether safeguards work, which age-appropriate design choices help, how AI literacy or adult support changes outcomes, and how healthy-use practices can be evaluated.
The fund is for independent research rather than a commercial product launch or a general-purpose business project. A proposal should make its research contribution clear. If it includes an intervention, tool, or design change, describe how the research will evaluate its effects and generate evidence others can use. Research focused on ChatGPT may receive preference, but the call is not limited to that product. The program also says it may favor interdisciplinary teams, underrepresented populations or regions, and studies that can inform near- or medium-term policy decisions.
Who should consider applying
Applicants must be at least 18. They must have an affiliation with a research institution or organization, and/or significant experience relevant to child or adolescent development, wellbeing, human-computer interaction, AI impacts, or a closely related field. The wording allows more than one credible applicant route; it does not say that every eligible person must hold a university appointment. Still, proposals need a team and institutional or organizational arrangements suitable for the work they promise, especially when they involve minors or sensitive information.
The opportunity is open to proposals from any country. OpenAI says it will prioritize countries with significant AI adoption, while also emphasizing inclusion and generalizability across populations, cultures, languages, and socioeconomic contexts. A non-U.S. team can apply, and a study need not be based in the United States. Multi-institution collaborations are allowed.
This call may suit university researchers, nonprofit research organizations, public-interest research teams, and experienced practitioners who can assemble an ethically responsible study team. It is less suitable for a for-profit initiative whose main purpose is to build or sell a product. OpenAI says it is seeking research rather than for-profit initiatives and will not prioritize for-profit organizations at this time. If a commercial partner contributes to a research project, make the independent research purpose, decision-making, data access, and conflict management explicit.
Research questions and approaches that fit
The official call accepts qualitative, quantitative, experimental, observational, theoretical, participatory, and mixed-methods work. Research with teenagers involved in designing or interpreting the study is welcome but not required. This flexibility should encourage applicants to choose a method because it answers the question responsibly, rather than because it sounds technically impressive.
A study might examine how the purposes of AI use relate to teens’ reported experiences; how family, school, culture, or access affects those experiences; whether an intervention changes a measurable outcome; or how patterns differ between developmental stages. Projects can also evaluate safeguards, parental or educator support, AI literacy, or product design. These are topic examples from the call, not prescribed study designs. Applicants should define their own question and explain why the chosen population, measures, comparison, and time period can answer it.
The proposal should add to existing evidence rather than simply repeat a broad claim that AI is beneficial or harmful. Identify the uncertainty the research will reduce. Define what “use,” “wellbeing,” “safety,” or “development” means in the study. If a project expects impacts to differ by age, language, background, or use pattern, specify how it will observe those differences and whether the planned sample can support them. If the work cannot support subgroup conclusions, say so and avoid promising more than the design can establish.
Funding amount and project expectations
The program has a total commitment of up to $5 million and may award up to $1 million to an individual research project. That ceiling is not a suggested request or a guarantee that a project will receive the full amount. Build a budget from the work plan: explain staff effort, participant-related costs, analysis, translation or accessibility needs, ethics and safeguarding work, and other major cost categories that apply. The application must identify the requested amount, major categories, and other committed support.
At most 10% of each grant may be used for overhead or indirect costs. Student, graduate student, and postdoctoral researcher involvement can be supported. Collaborations across institutions are permitted. The official announcement asks for a project plan and timeline, including key activities, milestones, and an expected completion date. OpenAI strongly prefers shorter projects that can produce final results sooner and asks for a working draft suitable for discussion at a small conference within 12 months after funds are received.
Awarded projects must provide an interim progress update in the first quarter of 2027, with early findings where appropriate. Each funded team must also provide a final report setting out its research question and methods, results and supporting evidence, limitations and sources of uncertainty, and recommendations for relevant stakeholders where supported by the results. Public release through a peer-reviewed publication, preprint, report, or similar channel is strongly encouraged, but publication is not a condition of receiving funding.
Eligibility and research ethics
Before preparing a full submission, check that the lead applicant is 18 or older and that the team can document relevant affiliation or experience. The funder’s eligibility language covers experience in child or adolescent development, wellbeing, human-computer interaction, AI impacts, and related fields. It does not list a required degree, country of residence, citizenship, or single discipline. Do not add requirements that the call does not state; instead, explain why the team’s expertise matches the methods and population.
Research involving minors requires careful planning. The application must explain the relevant institutional or independent ethics-review process; how informed consent and age-appropriate assent will work; how privacy and data will be protected; and how the team will respond to disclosures of harm or imminent risk. Applicants should also address whether they will collect sensitive personal, behavioral, or mental-health information, the team’s child-development, clinical, or safeguarding expertise, and parental consent where appropriate. OpenAI notes that sensitive data do not automatically make a proposal ineligible, but they may require additional review, expertise, safeguards, and time.
These are substantive parts of study feasibility, not paperwork to append after the research plan is complete. If participants could discuss distress, unsafe situations, or other sensitive topics, state who will review the protocol, who will handle a disclosure, what information will be collected, and how access and retention will be limited. Align the recruitment schedule with the ethics process and any required permissions. The official call does not prescribe one universal ethics approval route, so applicants should identify the process that applies to their own institution and jurisdiction.
Application materials
Applications should be prepared as a single-spaced Google document and submitted in English. The research itself may be carried out in any language. The announcement lists these components:
- Project summary: A brief overview of the proposed research and its expected contribution, under one page.
- Research proposal: Up to three pages, excluding references. Address the research questions, relevant prior research, methods and analysis, study population and recruitment strategy (including feasibility or access evidence), expected outputs, limitations, and relevance to the program’s goals.
- Project plan and timeline: Key activities, milestones, and expected completion date.
- Budget and justification: The funding request, major cost categories, and other committed sources of support.
- Project dependencies: Requirements that must be met to carry out the research and risks that could prevent completion.
- Ethics and safeguarding statement: Required for studies involving minors, sensitive information, or potentially vulnerable populations. Where relevant, refer to applicable regulations and the compliance plan.
- Team information: Short CVs or biographies for the principal investigator and key team members.
- Independence and conflicts disclosure: Relevant financial relationships, industry funding, and other potential conflicts.
The public announcement provides these requirements but does not publish a scoring rubric with numerical weights or a fixed award size for each topic. Use the linked official announcement and portal for the current submission form and any clarifications; do not assume an unpublished template, budget range, or institution-specific rule.
How proposals will be reviewed
OpenAI says applications will be reviewed by a panel of internal experts and advisors, on a rolling basis. Published criteria include scientific rigor, relevance to current public discourse on AI, proposal clarity and coherence, and the potential to translate findings into guidance for models, products, safeguards, regulation, or future research. The panel will also consider ethics and safeguarding, feasibility, team expertise, inclusion and generalizability, complementarity with OpenAI’s internal research agenda, and independence and credibility.
The independence criterion matters for a funder that also develops AI products. A credible proposal should show how its design can produce trustworthy findings even if results are unfavorable to a provider. Explain who controls the analysis, how contrary or null findings will be reported, and what limitations might prevent a strong conclusion. This is practical guidance drawn from the published criterion; the call does not prescribe a particular data-sharing or publication agreement beyond encouraging public dissemination.
To make the proposal easy to assess, keep the chain from question to evidence visible. For each aim, name the data or observations it requires, how those will be analyzed, and what result would count as support, no clear effect, or an unexpected outcome. Connect the output to a concrete audience without claiming that one study will settle a broad debate. An actionable result can be a careful estimate, a tested intervention, an account of differing experiences, or a clearly bounded recommendation.
Timeline and next steps
The application portal lists the deadline as October 6, 2026, at 11:59 p.m. PDT. The OpenAI announcement likewise says submissions remain open through October 6. Because the deadline is close, applicants should confirm that they can meet the stated formatting, ethics, budget, and disclosure requirements before spending time polishing language. Submit through the official application portal, not by email.
OpenAI says review will be rolling and selected proposals will be notified on or before November 13, 2026. The program’s public page does not provide a fixed project start date. Funded teams should plan for a progress update in Q1 2027 and a working draft within 12 months after receiving funds. The date funding begins is not confirmed in the public call.
A practical sequence for the remaining application time is to verify fit first, select one answerable question, check access to the study population and any required ethics review, and draft the proposal around the listed requirements. Then reconcile the schedule and budget, ask collaborators to confirm their roles and support, and conduct a final check for English submission, page limits, missing disclosures, and consistency between the proposed work and the requested amount.
Preparation strategy and common mistakes
Start with a narrowly framed research question that directly concerns teens’ AI use, its context, or a safeguard or design choice. A proposal becomes stronger when it identifies what is already known, what remains uncertain, and how the planned evidence will address that gap. Avoid stacking many topics into one short project. A study of one carefully justified population and outcome may produce more useful evidence than an unworkable promise to measure every dimension of teenage life.
Show feasibility with specifics. The call asks for evidence of access to the study population or a plausible recruitment strategy. Explain who can recruit participants, where the study will take place, how long data collection will take, and what happens if recruitment is slower than expected. For research with minors, do not assume access or permission: describe the ethics process, consent and assent, safety response, and protections for sensitive data.
Make the analysis match the claim. If the study is observational, avoid causal language unless the design supports it. If you intend to compare groups, explain how they will be defined and whether the sample supports that comparison. If a qualitative design is best suited to understanding experience, explain how interpretation will be documented and checked. State limitations as part of the plan rather than leaving reviewers to discover them.
The budget should map to the work rather than simply approach the maximum. Check the 10% cap for indirect costs, include only justified cost categories, and identify other committed support. The timeline should account for recruitment, ethics review, analysis, and the requested Q1 2027 update. A common weakness in proposals involving teenagers is an ethics section that names consent but not privacy, safeguarding, disclosure handling, or team expertise; the call asks applicants to address each relevant issue.
Finally, make independence tangible. The application should describe conflicts and financial relationships and explain how the team will report results with appropriate caveats. Do not design the study to produce a predetermined conclusion. OpenAI’s criteria expressly emphasize credibility regardless of whether findings favor AI product providers, and that standard is best met through transparent methods and proportionate claims.
Frequently asked questions
Is the program limited to U.S. researchers?
No. Proposals from any country are encouraged. The announcement says countries with significant AI adoption may receive priority. Research can be conducted in any language, though the application must be submitted in English.
What is the maximum grant?
OpenAI says it may award up to $1 million per individual research grant from a total program commitment of $5 million. The public announcement does not give a fixed grant amount for each project.
Can graduate students or postdoctoral researchers be paid from the grant?
Yes. OpenAI’s FAQ confirms grant funds may support student, graduate student, and postdoctoral involvement.
Does every project need to study ChatGPT?
No. The call concerns generative AI and teen development broadly. Preference may be given to research involving ChatGPT, but it is not stated as an eligibility requirement.
Must the project publish a paper?
Public dissemination is strongly encouraged through a peer-reviewed publication, preprint, public report, or similar format. OpenAI says publication is not a condition of receiving funding. Funded projects must still provide the required interim update and final report.
Can a commercial company apply?
The eligibility information does not frame the call as a for-profit business program. OpenAI says it is seeking research rather than for-profit initiatives and will not prioritize for-profit organizations at this time. Teams with commercial relationships should disclose them and show how independence and credibility will be maintained.
Where can applicants ask questions?
The program lists [email protected] for questions. It also says individual status updates will not be provided to applicants whose proposals are not selected.