Penn GSE’s Artificial Intelligence & Education (AIED) Lab, led by Professor Seiji Isotani, bridges artificial intelligence, learning sciences, and education policy to design AI-enabled technologies that are pedagogically grounded and responsive to real-world contexts—particularly in underserved and resource-constraint environments.
Advancing AI in Education
The AIED Lab, which serves Penn GSE's masters program in Learning Analytics and Artificial Intelligence, is where research and practice converge to solve the most pressing challenges in education. Our work is not siloed in theory; it is deeply intertwined with the pedagogical strategies and instructional designs we teach and implement in the program. By bringing real-world classroom experiences into the lab, we ensure that our AI solutions are not only technically advanced but also practically viable and human-centered. This cycle of continuous feedback between research and teaching empowers our members to see the immediate impact of their work, bridging the gap between innovative ideas and transformative learning outcomes.
About the Penn GSE Masters Program
Learning Analytics and Artificial Intelligence, M.S.Ed.—the Ivy League's first fully online master's program of its kind—leverages educational data and modern algorithms to develop learning technologies that improve student outcomes, support teachers and school leaders, and help identify and address gaps. Successful use of learning analytics requires a deep understanding of both algorithms and learning, including the social context in which learning occurs and learning analytics are applied.
Our Mission
The Three As: Agency, Access, and Augmentation
Inspired by the principles of Paulo Freire, our work is built on three core principles:
- Agency: Empowering a global citizenry to navigate an AI-driven world without compromising their humanity, critical voice, or autonomy
- Access: Designing inclusive AI solutions that bridge the digital divide, ensuring effectiveness across diverse environments and resource-constrained settings
- Augmentation: Leveraging AI as a powerful tool to amplify and enhance—not replace—human potential, scaling high-quality instruction to reach every learner
Our Partners
Relationships built on advancing AI in education.
Meet Our Lab Team
Lab Leadership
Seiji Isotani
Principal Investigator
Associate Professor
Faculty Director, Learning Analytics and Artificial Intelligence, M.S.Ed. Program
President of the International AI in Education Society
Nam Nguyen
Lab Manager
Senior Research Manager, AI in Education
Outreach and Community Manager, Learning Analytics and Artificial Intelligence, M.S.Ed. Program
Luis Gaitan
Research Associate & Teaching Fellow
Lead Developer at August Interactive
Visiting Researcher at Reitaku University
Research Affiliate at the Morality Lab, Boston College
Website: gaitaneluis.com
Shruti Mehta
Ph.D. Student
Doctoral Student, Learning Science and Technology, M.S.Ed. Program
Penn AI Fellow
Lab Members
Lab Alumni
In the Global South, where nearly 60% of the population still lacks reliable electricity, the AI revolution often feels like a distant promise, hindered by a "digital divide" that leaves millions of students behind. To bridge this gap, our team developed AIED Unplugged, a framework that bypasses infrastructure barriers by using mobile devices as proxies to provide high-quality, automated feedback on handwritten work—all without requiring a stable internet connection. The impact has been transformative: to date, the project has reached over 500,000 students across 7,000 schools in Brazil, slashing feedback wait times from four months to just 72 hours and prompting national policy shifts toward equitable AI. Today, at the AIED Lab, we are expanding this mission by designing new pedagogical strategies that ensure the most powerful educational tools remain human-centered and accessible to every learner, regardless of their connectivity.
As the rapid emergence of generative AI leaves many educators caught in the "EdTech Retrofitting Trap"—the friction of forcing high-performance tools into legacy 19th-century syllabi—the Penn AI Pedagogy Initiative (PAPI) offers a purpose-built architecture for the future of learning. Funded by the Penn AI Discovering the Future of AI grant and Penn GSE Dean’s Strategic Priority Grant, this initiative operates on a radical principle: pedagogy is the operating system, and AI is simply the processor. By pairing disciplinary experts with Penn GSE’s experts learning scientists in a model we call the "Synapse," we move beyond generic chatbots to co-design high-fidelity tools that target specific academic "stuck points" where students typically struggle. This collaborative engine is intended to scale across Penn’s 12 schools, transforming over 30 courses and empowering thousands of students to reclaim their intellectual agency.
In a world where artificial intelligence is often seen as a "black box" of impenetrable equations, our AIED Instructional Games initiative is turning the abstract into the accessible. The project confronts a critical educational divide: the challenge of teaching complex machine learning logic to students who lack both a stable internet and a foundational background in computer science. Drawing on the pioneering principles of AI Unplugged and CS Unplugged, we develop instructional games that translate invisible algorithms into tangible, kinesthetic exercises. This method has already enabled thousands of learners in resource-constrained regions to master high-level AI principles through hands-on play, fostering a more inclusive technical literacy.
Most of today’s generative AI is built on a narrow slice of human experience, trained predominantly on data from Western, Educated, Industrialized, Rich, and Democratic (WEIRD) populations—a bias that mirrors a long-standing problem in social science research itself. This over-representation means that while AI might excel in a Silicon Valley classroom, it often fails to account for the linguistic, cultural, and pedagogical nuances of the Global South. Our research addresses this "WEIRD" imbalance by investigating how Large Language Models perform in diverse settings—from rural India to urban China—to ensure that the next generation of AI tools respects and reflects the realities of the global majority. By evaluating these models through a non-Western lens, we are moving beyond a "one-size-fits-all" approach to create a truly inclusive digital future.
Contact Us
-
Seiji Isotani
-
Nam Nguyen