OpenAI is committing $5 million to assist unbiased research on how generative AI impacts the lives and development of younger folks ages 13–17.
Generative AI is changing into a part of how teenagers study, create, and search recommendation. Its results are prone to be nuanced. They might depend upon what teenagers use AI for, how they interact with it, their particular person circumstances, and the assist obtainable from households, faculties, communities, and expertise suppliers.
We need to construct a stronger proof base about the way in which teenagers use AI, in addition to the impacts of AI on their lives, to tell development of AI merchandise for teenagers and to assist reply main questions going through policymakers and regulators. For instance, we intention to deepen industry-wide understanding of the nuances throughout age and context, efficacy of safeguards and interventions, and the ways in which proof can inform implementation and analysis. This program will fund unbiased research into the alternatives and challenges related to AI use throughout adolescence, with a selected give attention to teenagers’ social and emotional development.
We welcome proposals that may deepen understanding and generate proof related to present and rising discourse, together with:
- How teenagers use generative AI and how completely different patterns of use relate to developmental outcomes
- Potential impacts of AI use in teenagers’ lives
- The particular person, social, cultural, and technical components that form the consequences of AI in teenagers’ lives
- Technical interventions, safeguards, and design decisions that may improve advantages or preserve teenagers safer
- How results might differ throughout ages, developmental levels, populations, cultures, languages, and socioeconomic contexts
We welcome interdisciplinary research involving fields akin to youngster and adolescent development, psychology, public well being, human-computer interplay, communication, sociology, anthropology, information science, and AI security. We encourage proposals from any nation, and will prioritize these from nations with vital AI adoption.
Qualitative, quantitative, experimental, observational, theoretical, participatory, and mixed-methods approaches are all eligible. Research involving teenagers within the design or interpretation of research is welcome however not required.
We encourage submissions that relate to a number of of the next subjects:
- Emotional development and well-being: Research on emotion regulation, resilience, id, company, vanity, and social connection.
- Social development and relationships: Research on friendships, household relationships, belonging, communication, empathy, social abilities, and the connection between AI and human connection.
- Patterns, contexts, and variance throughout demographics: Research on how outcomes range by time spent, kind and goal of use, age, tradition, language, socioeconomic circumstances, and different particular person or contextual components.
- Mitigations, age-appropriate design, and coverage approaches: Research evaluating safeguards, product interventions, AI literacy, parental or educator assist, and healthy-use practices.
Research involving minors requires specific consideration to knowledgeable consent and assent, privateness, safeguarding, information safety, and the likelihood that participation might expose or generate delicate data.
Applicants proposing research involving minors should clarify:
- The related institutional or unbiased ethics-review course of
- How knowledgeable consent and age-appropriate assent shall be dealt with
- How individuals’ privateness and information shall be protected
- How researchers will reply to disclosures of hurt or imminent danger
- Whether the challenge includes delicate private, behavioral, or mental-health data
- What related child-development, medical, or safeguarding experience the research staff possesses
- Parental consent, the place applicable
Collecting delicate data doesn’t robotically make a challenge ineligible. However, such initiatives might require extra moral evaluate, experience, safeguards, and time.
Applications ought to be formatted as a single-spaced Google doc. Applicants ought to submit:
- Project abstract: A quick overview of the proposed research and its anticipated contribution, lower than one web page.
- Research proposal: No greater than three pages, excluding references, addressing:
- Research questions
- Relevant prior research
- Proposed strategies and evaluation
- Study inhabitants and recruitment technique, together with proof of feasibility or entry
- Expected outputs
- Potential limitations
- Relevance to this system’s objectives
- Project plan and timeline: Key actions, milestones, and anticipated completion date.
- Budget and price range justification: Requested funding, main price classes, and some other dedicated sources of assist.
- Project dependencies: An rationalization of necessities that have to be met for the research to be carried out and dangers that may endanger the challenge’s fruition.
- Ethics and safeguarding assertion: Required for any research involving minors, delicate data, or probably susceptible populations. Where related, proposals ought to explicitly reference any regulation governing research with teenagers and compliance technique.
- Team data: Short CVs or bios for the principal investigator and key staff members.
- Independence and conflicts disclosure: Relevant monetary relationships, {industry} funding, or different potential conflicts of curiosity.
Applications have to be submitted in English. Research itself could also be carried out in any language.
Proposals shall be evaluated based mostly on:
- Scientific rigor: The power and appropriateness of the research design and evaluation
- Relevance: Examination of a subject related to present public discourse on AI
- Proposal high quality: The readability, focus, and coherence of the applying
- Actionability: The potential to translate findings into steering for fashions, merchandise, safeguards, regulatory approaches, or future research
- Ethics and safeguarding: The adequacy of protections for minors and different individuals
- Feasibility: Whether the challenge may be accomplished with the proposed sources and timeline
- Team experience: The staff’s skill to conduct the proposed work responsibly
- Inclusion and generalizability: Attention to various populations, contexts, cultures, and experiences
- Complementarity: The extent to which the work provides to present proof and enhances OpenAI’s inner research agenda
- Independence and credibility: The challenge’s skill to supply reliable findings no matter whether or not these findings are favorable to suppliers of AI merchandise
Preference could also be given to research involving ChatGPT, interdisciplinary groups, underrepresented populations or areas, and initiatives able to producing clear proof that may inform near- or medium-term coverage and regulatory choices.
Final grants could also be topic to the grantee’s execution of a grant settlement, award letter, certification, acknowledgment, or different documentation setting forth relevant phrases and circumstances. Grantees should settle for and adjust to any such necessities as a situation of receiving grant funding.
Submissions are open at this time by October 6, 2026. A panel of inner consultants and advisors will evaluate functions on a rolling foundation and notify chosen proposals on or earlier than November 13, 2026. Follow this link(opens in a new window) to use.
Q: What are the anticipated outputs?
A: Every funded challenge shall be anticipated to offer OpenAI with a ultimate report describing:
- The research query and strategies
- Results and supporting proof
- Limitations and sources of uncertainty
- Recommendations for policymakers, regulators, expertise suppliers, or different stakeholders, the place supported by the findings
To guarantee research is progressing, we require an interim replace on research progress and, the place applicable, a sign of early research findings in Q1 2027.
We are prioritizing research that may advance public dialogue on these essential subjects. We strongly encourage researchers to make their findings publicly obtainable by a peer-reviewed publication, preprint, public report, or related format. However, publication is not going to be a situation of receiving funding.
Q: What are the eligibility standards?
A: To be eligible, candidates have to be:
- 18 or older
- Affiliated with a research establishment or group, and/or vital expertise related to youngster or adolescent development, well-being, human-computer interplay, AI impacts, or a intently associated discipline
Note that we’re in search of to fund research slightly than for-profit initiatives, so we is not going to prioritize for-profit organizations right now.
Q: What price range is obtainable per challenge?
A: We will award particular person research grants as much as $1 million and totaling as much as $5 million.
Q: Can grant funds be used to cowl overhead or oblique prices?
A: A most of 10% of every challenge’s grant can assist overhead or oblique prices.
Q: Can grant funds be used to assist pupil, graduate pupil, or post-doc involvement in research?
Q: Are these grants funded by OpenAI Foundation or OpenAI Group PBC?
A: The grants are funded and administered by OpenAI Group PBC.
Q: Who can I contact for questions on this grant?
A: Email collaborativeresearch@openai.com. Note that we’ll not be offering standing updates for particular person functions that aren’t chosen.
