Sharing the safeguards, policies, and experts we work with to guide our approach to teen use of safe AI.
Teens are the first generation growing up with AI, and this technology will heavily shape their future. Today, nearly 9 in 10 teens on ChatGPT use it for learning, information, skill-building, or productivity in a single week. This is why we believe it’s critical for teens to have access to AI. Keeping teens from using it until adulthood would be like asking a previous generation to avoid the internet or search engines until they turned 18, leaving them less prepared to use one of the defining technologies of their time. But access must be paired with protections designed specifically for teens.
During a recent OpenAI Academy(opens in a new window) Oriana McKenzie shared how ChatGPT helped her become No.1 in her high school class by studying smarter and managing her workload—giving her the time needed for the sports and school programs she loves. Her story illustrates what’s possible when teens have access to AI that helps them learn and create.
To maximize those gains, teens need stronger protections designed for their stage of life, including automated guardrails that let them explore, learn, and build with confidence, while providing safeguards tailored to their age. As AI becomes more capable, our responsibility is to combine broad access with age-appropriate protections that allow teens to benefit from this technology as it evolves.
Over the past year, that has meant strengthening default protections for teens, rolling out age prediction, expanding Parental Controls, creating additional family resources(opens in a new window) to help parents support healthy, responsible use, and introducing learning features to help support deeper understanding rather than just providing answers.
Our work is guided by four key commitments:
- We put teen safety first even when it may conflict with other goals.
- We encourage real-world support in times of need.
- We treat teens as teens.
- We are transparent by setting clear expectations.
Learning is one of the clearest ways teens benefit from AI, so we’re building tools that encourage active engagement, critical thinking, and deeper understanding. Study Mode(opens in a new window) was designed in collaboration with teachers, learning scientists, and pedagogy experts to help students work through problems step by step using guiding questions, structured explanations, and opportunities for reflection without simply providing the answer. Early evaluation of tools like Study Mode has shown promising gains in student performance, helping inform our broader research into how AI can support learning outcomes.
While teens can turn this mode on themselves, now parents with linked teen accounts can turn on Study Mode directly from Parental Controls. When enabled, it is on by default whenever a teen starts a new chat, giving families another way to guide how ChatGPT is used for schoolwork and study.
We also recently introduced education-focused starter prompts for teens so it is easier to begin with tasks like breaking down a topic into simple steps, turning notes into a study guide, creating flashcards or practice questions, and checking evidence and clarity.
We’re continuing to expand interactive learning experiences as well. Research has consistently shown that people learn more effectively when they can actively engage with concepts rather than passively consume information. Since launching earlier this year, 18 million weekly users now engage with interactive math and science experiences in ChatGPT, and we’ve expanded those experiences to more than 300 topics total from integrals and mitosis to moon phases, photosynthesis and more. We also introduced a pronunciation experience that uses audio to help people learn how to pronounce words in more than 61 languages.
If our system estimates someone using ChatGPT is under 18, we automatically provide a more age-appropriate experience(opens in a new window). Teens can still use ChatGPT to learn, create, and explore, but with additional protections designed to reduce exposure to content that may not be appropriate for them. That includes stronger safeguards around graphic violence, self-harm, risky viral challenges, unhealthy body-image content, and dangerous, romantic, or sexual roleplay. The North Star is to help ensure ChatGPT remains a tool for learning and creativity and not a substitute for real-world relationships.
We’re also designing ChatGPT to encourage healthy habits. Teens who spend extended time in ChatGPT will now receive more frequent break reminders that encourage them to pause and step away. These reminders are intended to help young people build balanced technology habits while still giving them the freedom to use AI for learning, creativity, and problem-solving.
We believe the best safety tools can help families navigate teens’ online experiences together. We built parental controls and notifications that allow them to set quiet hours, turn off voice mode, manage access to image generation, and receive notifications in certain high-risk situations, such as indications of potential self-harm.
We’re expanding those notifications to include cases where a linked teen account has been deactivated for violating our usage policies on violent threats or acts of violence online. This approach helps parents know when something serious has happened while respecting teens’ privacy and encouraging conversations and support offline.
As we developed these features we consulted with several experts, including Moonshot(opens in a new window), a leader in preventing online violence, to help ensure these interventions are thoughtful, effective, and focused on helping families respond in times of need.
We also want to ensure parents, teens, and educators have clear guidance on how to best use AI responsibly. We publish educational resources in more than 50 languages, including A Family Guide to Help Teens Use AI Responsibly(opens in a new window), Tips for talking to your teen about AI(opens in a new window), and Tips for parents using ChatGPT(opens in a new window).
Keeping young people safe online isn’t something any one company can solve alone. We work closely with teens, parents, educators, child safety experts, mental health professionals, researchers, governments, and civil society organizations. These partnerships help us better understand the needs of young people, identify emerging risks, evaluate the effectiveness of our safeguards, and ensure our policies are informed by independent expertise.
This work spans our approach to both learning and youth wellbeing. We collaborate with educators, including the American Federation of Teachers, students, and learning organizations to better understand how AI can support teaching and learning while reinforcing active engagement and critical thinking. We also work closely with experts across the mental health, child development, and youth safety communities, including the American Psychological Association(opens in a new window), our Global Physicians Network, and our Expert Council on Well-Being and AI.
Our work to build a healthier digital ecosystem extends beyond our own products. We’ve called for developing scalable global standards for youth AI safety through a dedicated AI Safety Institute, open-sourced policies and tools that developers can build into their own AI systems, and work with organizations including Common Sense Media(opens in a new window) and everyone.ai(opens in a new window) to share learnings and best practices.
This year, we also joined the Family Online Safety Institute(opens in a new window) (FOSI), a long-standing advocate for safe online experiences for families and young people. We will work together to ensure young people can benefit from AI, with empowered parents and responsible industry practice at the center. We hope this partnership will help advance shared understanding of what works, strengthen age-appropriate protections and parental tools, and support industry-wide efforts that enable young people to benefit from AI safely and responsibly.
Protecting young people online requires continuous progress. In the coming months, we’ll continue strengthening age-appropriate protections, giving parents more tools and control, improving safeguards against serious harms, advancing research on healthy AI use, and building more experiences that help teens learn actively and use AI with confidence.
We know there is more to do, and we’re committed to continuing our work in collaboration with teens, parents, educators, experts, and communities worldwide to bring safe broad access to teens.
Facts Only
* Nearly 9 in 10 teens on ChatGPT use it for learning, information, skill-building, or productivity in a single week.
* Access to AI is advocated for teens as they are the first generation growing up with the technology.
* The organization has implemented changes including strengthening default protections for teens and rolling out age prediction over the past year.
* Features introduced include Study Mode, which uses guiding questions for step-by-step problem-solving.
* Education-focused starter prompts were introduced to aid in tasks like creating study guides or flashcards.
* Interactive learning experiences have expanded across over 300 topics in math and science, plus a pronunciation experience for over 61 languages.
* Systems estimate usage under 18 and automatically provide a more age-appropriate experience with added protections against exposure to content like self-harm or risky roleplay.
* Parents can activate Study Mode via Parental Controls when linked teen accounts are active.
* Parental controls allow setting quiet hours, managing image generation access, and receiving notifications for high-risk situations.
* The organization consults with experts such as Moonshot to inform safety interventions.
Executive Summary
Access to AI is argued for teens because they are the first generation growing up with this technology and will be heavily shaped by it, with nearly 9 in 10 teens using ChatGPT for learning or productivity weekly. The argument posits that restricting access until adulthood would leave teens unprepared for a defining technology of their time, drawing an analogy to restricting access to the internet or search engines. This access must be paired with specific protections tailored to teenagers.
To maximize benefits, safeguards are necessary, including automated guardrails and age-appropriate protections based on the teen's stage of life. The organization has implemented several protective measures over the past year, such as strengthening default protections, rolling out age prediction, expanding Parental Controls, creating family resources, and introducing learning features designed to promote understanding rather than just providing answers.
The approach centers on four commitments: prioritizing teen safety, encouraging real-world support, treating teens as teens, and ensuring transparency through clear expectations. Learning tools, like Study Mode, were developed with educators to encourage active engagement and critical thinking. Furthermore, interactive learning experiences have been expanded across various subjects, and age estimation automatically provides a more appropriate experience, imposing safeguards against exposure to harmful content like graphic violence or risky roleplay. Parental controls allow families to manage usage and receive notifications regarding high-risk situations, often in coordination with experts like Moonshot.
Full Take
The narrative frames the debate around a tension between granting broad access to a powerful tool and ensuring developmental safety. The core pattern involves reframing technological adolescence—the transition into adulthood—as a critical period demanding specific, layered protections rather than simple age gating. The organization attempts to bridge this gap by framing AI interaction not just as consumption but as a mechanism for active learning and critical engagement, exemplified by Study Mode.
The underlying assumption is that knowledge acquisition through interactive processes is superior to passive information reception, which supports the push for learning-focused features. This connects to a broader pattern where entities seeking to control an evolving domain often resort to defining specific parameters (guardrails) rather than outright prohibition, aiming to shape the trajectory of adoption. The implementation of parental controls and external expert consultation suggests a systemic acknowledgment that this is not solely a product design issue but an ecosystem responsibility requiring triangulation across safety science, pedagogy, and family dynamics.
The move toward collaborative standards—seeking global AI safety institutes and open-sourcing policies—suggests a recognition that individual corporate responsibility is insufficient against emergent technology risks. The pattern here is the institutionalization of external accountability; leveraging partnerships with child safety experts and educators functions to shift the locus of control from pure platform governance to a shared social responsibility, attempting to establish a resilient boundary between innovation and harm.
Bridge Questions: If access to AI fundamentally changes cognitive development, what metrics should govern the balance between exploratory freedom and mandated safety boundaries? How can the commitment to real-world support be operationalized across diverse cultural and educational contexts when developing global standards? What is the long-term risk if perceived privacy or safety tools are implemented in a way that unintentionally stifles the very critical thinking AI is designed to foster?
Sentinel — Human
The text reads as a well-structured public statement from an organization detailing efforts to balance AI access with safety, supported by specific examples and collaborative frameworks.