Code.org Shifts Strategy as AI Reshapes What Schools Should Teach
Code.org, the nonprofit focused on expanding computer science access in K-12 schools, is broadening its mission beyond coding instruction to include AI literacy and teacher training. The organization's new CEO, Karim Meghji, said the rise of AI makes computer science education more critical, not less.
"Computer science is going to change," Meghji said. "Every subject is going to change." He rejected the notion that AI signals the decline of computer science, arguing instead that the two fields have a complicated relationship: computer science created the foundation for AI, but AI will fundamentally alter how computer science is taught.
What Students Actually Need to Know
Meghji drew a parallel to biology education. Students dissect frogs in middle school not to train future biologists, but to understand anatomy. Similarly, he said, students should learn how AI models work-not just how to use them.
He identified three foundational concepts every student should grasp:
- How AI models are designed and trained
- The role of data in shaping model outputs and introducing bias
- How these systems can produce unreliable answers or hallucinations
Understanding these basics would give students the ability to use AI systems more effectively and critically, regardless of their future career path.
Schools Are Adopting AI Unevenly
Adoption remains scattered across districts. Some teachers and schools are experimenting with AI-adapted curricula. Others are trying to fold AI into existing lessons without fundamentally rethinking how they teach. Still others have adopted policies against classroom AI use.
Meghji described the current state as "the messy middle"-a phase that may last longer than typical technology rollouts. Districts that understand AI's significance are developing policies, while others have yet to engage with the technology.
Code.org is shifting its policy advocacy from pushing computer science graduation requirements to enhancing those requirements with AI literacy components.
What Vendors Get Wrong
Education companies racing to bring AI tools to classrooms often focus on the wrong things, Meghji said. Most emphasize how to use AI tools rather than how they work. This approach reduces critical thinking and creates passive users.
He also flagged a practical barrier: AI models are expensive to run. Vendors charge for API access, which prevents schools from offering free, scaled access to students. Meghji called on vendors to provide free or subsidized credits to educational organizations.
Safety testing matters too. Code.org plans to pilot tools with small groups before broader rollout, ensuring classroom safety before market-scale deployment. Trust with educators and students, he said, is paramount.
Code.org's Next Phase
Beyond curriculum development, the organization plans to expand into teacher training and policy work. Most K-8 teachers lack deep technical backgrounds in computer science or AI, creating a gap between what students need and what teachers feel equipped to teach.
Meghji also signaled plans to expand internationally. After establishing a strong presence in the U.S., Code.org is considering India as its next primary market.
Internally, he emphasized the need for organizational adaptability. Education companies cannot serve students and teachers effectively if their own teams aren't prepared for an AI-integrated world. Culture and mindset matter as much as product development.
Learn more about AI for Education and explore resources for AI Learning Path for Teachers.
Your membership also unlocks:
