Five years ago, a mere handful of universities ran a summer program built specifically around artificial intelligence. Today, more than a dozen do, and new options keep appearing. Artificial intelligence programs for high schoolers have multiplied rapidly, with the strongest among them offering something a classroom rarely can: the chance to build, question, and evaluate AI systems alongside university faculty and motivated peers. The challenge is no longer finding an option. It is telling the rigorous programs apart from the ones that offer little more than a certificate.
This guide looks at the top 15 artificial intelligence programs for high school students, with verified details on cost, eligibility, selectivity, and college credit. Just as importantly, it offers a framework for judging quality — because in a field where society is still learning what responsible use looks like, the best programs teach students to question AI, not just to use it.
Artificial intelligence isn’t just a specialty subject any more.
It shapes medicine, law, finance, agriculture, and the arts, and universities increasingly expect incoming students to reason about it across disciplines. AI programs for high school students give motivated learners a structured way to move beyond headlines and classroom demonstrations into real technical and analytical work.
The strongest programs do two things at once. They build concrete skills, including Python, model training, data analysis, and policy reasoning, and they cultivate judgment about when and how AI should be used. As AI tools become universal, the differentiating skill is no longer access to the technology but the ability to appraise its ethical and responsible use. Students who can explain not only how machine learning models work but also where they fail, and who bears the consequences, are prepared for universities and careers where that judgment and decision making will matter.
There is also a practical consideration. AI courses for high school students range from free, self-paced modules to selective residential institutes costing more than $9,000. The difference in lasting value rarely tracks the price tag. It tracks how much genuine work the student does, and how seriously the program oversees that work. If deep learning is a priority, for example, confirm that the curriculum actually covers it rather than assuming every AI program does. (If you’re looking for shorter, camp-style experiences, see our guide to the top artificial intelligence camps for high school students.)
The highest quality AI programs for high school students fall into two categories:
The strongest programs tend to share two qualities: a high degree of student agency in shaping the work, and meaningful institutional oversight that ensures academic rigor. Programs with oversight from universities and established non-profits typically possess these qualities, giving students the closest approximation to what college-level work actually looks like. That kind of academic structure can prepare students for more advanced study.
To read more about the different types of programs available, and how they differ from popular “independent mentor-matching” services, read our comprehensive guide on choosing the best research opportunities.
One additional consideration applies to AI specifically. Because the field’s ethical questions are still unsettled, programs that examine AI from multiple disciplinary angles, including technical, economic, and social perspectives, position students to think critically about responsible AI rather than simply absorb its tools. Look for programs with verifiable standards and oversight, and for evidence that students do substantial work of their own.
The programs below were selected based on:
Carnegie Mellon’s AI Scholars is a fully funded four-week program for rising high school seniors. Participants take college-level coursework in artificial intelligence and machine learning, receive mentorship from CMU faculty and graduate students, and complete collaborative projects that apply AI concepts to tackle real world challenges and culminate in a final symposium presentation. The technical curriculum is supplemented by weekly college-readiness seminars covering topics such as admissions, financial aid, academic preparation, and student wellbeing.
For a closer look at the program, including application requirements and whether it is worth it, read our complete guide to the CMU AI Scholars program.
The Global Problem-Solving Institute is the rare AI program built on the premise that understanding artificial intelligence requires more than code. In the Analyzing AI problem area, international teams of five to six students work under university professors to examine how AI reshapes economies, societies, and human judgment, then design and defend an original solution project. The structure is deliberate: professor-led instruction, design-thinking and systems-thinking methods, and accredited academic oversight through UNC Chapel Hill.
According to Pioneer Academics, students are assessed before and after the program across 44 innovation behaviors based on a research framework developed at UVA’s Darden School of Business — documented skill growth, not just a certificate. High school students of any age can apply, and the program is built with 9th and 10th graders specifically in mind, making it one of the few options offering substantive engagement with AI’s hardest questions this early in high school.
RSI, run by the Center for Excellence in Education in collaboration with MIT, is a cost-free, six-week residential STEM research program for students who have completed the equivalent of 11th grade. Participants begin with intensive STEM coursework, followed by a five-week research internship in which they conduct an individual project under an experienced scientist or researcher. Research placements span numerous STEM disciplines and have included machine learning and other computational fields. During the final week, students prepare written and oral presentations of their research.
To help decide whether RSI is right for you, you can check out our comprehensive program guide.
Beaver Works, a collaboration between MIT Lincoln Laboratory and the MIT School of Engineering, offers a four-week, project-based summer program for rising high school seniors. Its CogWorks: Autonomous Cognitive Assistant course focuses on machine learning and data science, with published curricula covering audio processing, computer vision, natural language processing, and neural networks.
Students work in teams on milestone projects that build toward an end-to-end cognitive assistant and final challenge. Before applying to the summer institute, students complete online prerequisite coursework designed to prepare them for the program’s technical demands.
For a fuller picture of the Beaver Works program, we encourage you to read our full guide.
Princeton AI4ALL combines technical AI instruction with an explicit focus on ethics, societal impact, and public policy. Students attend lectures from Princeton faculty, learn core machine-learning concepts, and complete small-group research projects led by graduate-student instructors and focused on applications of AI for social good.
The three-week residential program also includes a trip to Washington, D.C. exploring the intersection of AI and public policy. The program is open specifically to low-income rising 11th graders, is provided at no cost to participants, and does not require prior programming experience.
Stanford AI4ALL introduces 9th graders to artificial intelligence through faculty lectures, live demonstrations, and small-group research projects led by Stanford graduate students and postdocs in computer science and AI.
Students are placed in one of four research project groups — Computer Vision, Medical AI, Natural Language Processing, or Robotics — while the broader curriculum examines how AI can address challenges in areas such as healthcare, disaster response, and poverty. No prior AI or programming experience is required.
The MIT Jameel Clinic bootcamp compresses a focused education in health AI into a single intensive week. Students receive daily instruction from MIT faculty and researchers in machine learning for health, clinical AI, drug discovery, and Python programming, with Python instruction divided by experience level.
They also work in groups on projects that culminate in a final presentation evaluated by program instructors. For students drawn to medicine as much as machine learning, the program offers direct exposure to how AI is being applied to clinical care and drug discovery, including the risks and challenges involved in deploying these systems.
Stanford’s AIMI internship gives high school students two weeks of virtual, mentored research at the intersection of artificial intelligence and medicine. Students build technical skills, work with real-world health data, and complete collaborative projects with guidance from Stanford student leads and researchers. Summer-program alumni may also be invited to continue into AIMI’s Academic Year Research Internship, an extended mentored program in which students pursue health-AI research in small teams.
COSMOS, the California State Summer School for Mathematics and Science, places students in faculty-led clusters for four weeks of intensive study. UC San Diego’s Machine Learning: Can We Teach a Computer to Think? cluster introduces students to Python, machine-learning algorithms and packages, and hands-on applications across several fields. Students complete projects throughout the cluster and develop a final COSMOS project that they present at the end of the program. The major eligibility constraint is geographic: for 2026, COSMOS was limited to California residents.
For full details on eligibility and the application, read our complete COSMOS participation guide.
ESAP’s Artificial Intelligence course combines practical generative-AI tools with the mathematics underlying modern machine learning. Students study calculus and linear algebra, vectors and matrices, neural networks, stochastic gradient descent, backpropagation, and several model architectures before spending the final week on a small-team AI project that culminates in a presentation and evaluation.
The three-week residential program at Penn Engineering combines faculty-led instruction with hands-on project work and awards 0.5 Penn Credit Units upon successful completion.
Our ESAP participation guide covers the application process and what to expect.
Wharton’s AI Leadership program approaches artificial intelligence from the business and leadership side, examining how organizations use AI, where it can create value, and the ethical and governance issues that come with deployment.
Students learn from Wharton faculty and industry experts, work in teams on a capstone that applies AI tools to a real-world business challenge, and participate in site visits to AI and technology companies in the Bay Area. For students whose interest in AI runs toward strategy, innovation, and entrepreneurship rather than technical research, the program offers a distinctly business-focused pathway into the field.
Georgetown’s AI Academy is a compact introduction to artificial intelligence with a distinctive Washington, D.C. perspective. Students examine AI’s scientific foundations and limitations alongside its ethical, social, regulatory, and geopolitical implications.
Lectures, guest speakers, guided discussions, hands-on exercises, and off-site visits give students opportunities to experiment with AI tools and simple models while considering issues such as bias, transparency, privacy, security, and accountability. The broad curriculum makes it an accessible introduction for students who are still exploring which aspects of AI interest them most.
Yale Young Global Scholars brings together students from more than 100 countries, and its Innovations in Science & Technology track gives AI-interested students an interdisciplinary STEM curriculum taught through seminars, lectures, and team capstone projects. AI is one thread within a broader scientific program rather than the sole focus — a strength for students who want to situate machine learning within wider scientific and global contexts.
We’ve written a full guide to Yale Young Global Scholars, including application strategy and cost considerations.
Cornell Precollege Studies places high school students directly into regular Cornell undergraduate courses taught by Cornell instructors, with letter grades and college credit recorded on an official Cornell University transcript. AI-related offerings vary from year to year; recent online options have included a course on how large language models work, their potential, and their limitations, as well as coursework incorporating generative AI. Students take the same condensed university courses and are held to the same academic standards as undergraduate students, making this college coursework rather than a dedicated AI research or enrichment program.
Our complete guide to the Cornell Precollege residential program walks through courses, costs, and logistics.
The Center for AI Safety’s virtual course examines what can go wrong as AI systems become more capable and how technical safeguards, institutions, incentives, and governance can reduce those risks. The course requires about five hours per week over eight weeks of readings, lectures, assignments, and facilitated discussions, followed by a four-week personal project phase.
It requires no prior technical background and draws participants from a range of academic and professional disciplines. However, it is not specifically designed as a high school program, and the current course materials do not explicitly confirm eligibility for high school students.
Start with the taxonomy above, but look closely at how each individual program is structured. University-driven programs may offer institutional oversight, while non-profit and industry-driven programs can offer access to specialized researchers and real-world problems. More important than the sponsor type is the actual academic structure: who supervises the work, how students are evaluated, and whether they develop their own question or contribute to a predefined project.
Then apply the harder test: how much of the work will actually be yours? A program where students follow a polished tutorial produces a polished tutorial result. A program where students must frame a question, defend an approach, make substantive decisions, and present original work asks something different of them. Look at whether students simply follow instructions or are expected to evaluate an AI model, make their own decisions, and solve real world problems. Those experiences can develop critical thinking, intellectual initiative, and the ability to engage with complex questions, qualities that selective universities such as Yale explicitly emphasize in their admissions guidance.
That doesn’t mean students should choose a program simply because they think it will strengthen their college applications. The academic experience itself matters more: what students learn, how much responsibility they take for the work, and whether the program gives them a foundation they can build on later.
Comfort with the subject material matters too. Many research-intensive AI programs target students in the later years of high school, although substantive options for younger students also exist. The Global Problem-Solving Institute is open to students in grades 9–12 and is specifically promoted by Pioneer as an option for 9th and 10th graders ready for university-level, professor-led work. Students who successfully complete GPSI with the required grade earn two college credits from the University of North Carolina at Chapel Hill.
More broadly, Pioneer Academics emphasizes formal academic standards and institutional oversight, including accreditation through Oberlin College. If you want to understand how Pioneer Academics approaches academic standards and oversight across its programs, this overview is the place to start.
The best AI program depends on what you want from the experience. Programs such as Carnegie Mellon’s AI Scholars, Pioneer’s Global Problem-Solving Institute (Analyzing AI), MIT’s Research Science Institute, Princeton and Stanford AI4ALL, and MIT Beaver Works offer very different combinations of coursework, research, project work, and mentorship. CMU AI Scholars and Princeton AI4ALL are fully funded but restricted to rising seniors and rising juniors respectively, while GPSI is open to students in grades 9–12 and awards two college credits from UNC-Chapel Hill to students who successfully complete the program with the required grade.
Evaluate programs based on the actual curriculum, mentorship, academic oversight, student ownership of the work, and the final output students are expected to produce.
Yes. Some major research programs target older high school students, but several substantive options are available to younger students. Stanford AI4ALL admits current 9th graders, Wharton’s AI Leadership accepts students in grades 9–11, and Pioneer’s Global Problem-Solving Institute is open to students in grades 9–12.
GPSI students work in teams with guidance from university professors and can earn two college credits from UNC-Chapel Hill. Starting early can give students more time to build AI literacy, technical skills, and experience thinking through the ethical and social questions surrounding the technology.
Some do, although most of the programs in this guide do not. Three currently offer university credit: Pioneer’s Global Problem-Solving Institute awards two credits from UNC-Chapel Hill to students who meet the completion standard; Penn’s Engineering Summer Academy awards 0.5 Penn Credit Units; and Cornell Precollege Studies places grades and regular Cornell undergraduate credits on an official Cornell transcript.
College credit can provide an additional form of academic validation because the work is formally assessed for university credit, but it should be considered alongside the curriculum, instructors, mentorship, and standards of evaluation.
A 2023 investigation by ProPublica and The Chronicle of Higher Education examined paid online research programs for high school students and documented questionable practices at some providers, including publication-focused marketing, questionable journals and preprint platforms, misleading descriptions of mentor credentials, and heavy mentor involvement in student work.
The investigation’s scope was specific to the growing paid research industry, but its questions are useful when evaluating any program. Ask who formally oversees and evaluates the academic work, whether the program’s standards are documented and verifiable, what role mentors play in producing the final project, and whether publication or another impressive-looking outcome is being effectively guaranteed.
Costs vary widely. Several programs in this guide, including CMU AI Scholars, Princeton AI4ALL, RSI, and MIT Beaver Works, are free for admitted students. Paid options range from roughly $3,100 for Pioneer’s virtual Global Problem-Solving Institute to $9,800 for Stanford AI4ALL’s residential program, while credit-bearing programs such as Cornell Precollege Studies are priced by the number of credits taken and can cost more. Price is not a reliable measure of program quality or selectivity.
Many paid programs offer need-based financial aid, although eligibility and award amounts vary. GPSI’s $3,100 tuition, for example, includes the cost of two UNC-Chapel Hill credits, and Pioneer states that it meets demonstrated need for admitted students when that need can be reliably verified.
Artificial intelligence will be part of whatever field a student eventually chooses, which makes how they first learn it consequential. A program can hand students tools, or it can teach them to question those tools — and the second kind of learning is what lasts. The strongest opportunities are the ones that match a student’s goals, push them intellectually, and result in real learning rather than just a title or a certificate.
High school students interested in conducting the highest level of research for high school students should consider joining a Pioneer information session to learn more about the Pioneer Research Institute.
Based on a recent survey, 71 percent of Pioneer Research scholars’ acceptances were to the top 20 US colleges and universities. Additionally, our alumni report acceptances to highly-selective institutions at a rate five times higher than the school’s published acceptance rate.
If you are a 9th or 10th grader, we encourage you to also check out the Global Problem-Solving Institute. You’ll have the rare opportunity to take an interdisciplinary approach to complex world programs, while earning college credits from UNC-Chapel Hill.
Doing research is commonplace. How do you choose the research opportunity that makes a difference?
Join us for a free online info session to learn about Pioneer
Check exclusive sharings From directors of prestigious programs
Sign up to receive alerts for upcoming deadlines and other opportunities for high-achieving students.
