This is Part 7 of my series on the 2025 AI Index. We've talked about models, performance, and policy. Now it's time to turn to the classroom.
The AI Index is a rich and comprehensive report, but like in previous years, the education chapter does not match the scale or urgency of the topic. There is less content here than the moment deserves. I will likely do a deeper dive on it soon.
Still, this is a good opportunity to look closely at two things: the growing use cases for AI in education, and the critical need to prepare future generations not just to use these tools, but to understand and shape them.
We should not repeat the mistakes made with social media. Platforms shaped a generation before schools, parents, or policymakers had a chance to prepare them. In many ways, we still have not, and young people are paying the price.
1. AI for Education: Where the Technology Can Help and How
Artificial intelligence has the potential to improve many aspects of education, from everyday classroom experiences to nationwide curriculum design. While actual deployment is still limited, the range of possible applications is limitless, and it is worth understanding where this technology can add real value. Here is a non-exhaustive list of AI use cases in education:
Supporting Students Directly
- Personalized learning: AI can help tailor lessons, exercises, and feedback to each student's level and pace. This could be especially useful for students who are either struggling or far ahead of their peers
- Tutoring and explanation: Large language models can answer student questions in plain language, offer multiple explanations for a concept, or guide them through problem-solving steps
- Language support: AI translation and transcription tools can make learning materials more accessible to students who speak different languages
- Accessibility: AI can assist students with disabilities through features like real-time captions, text-to-speech, or summarization of complex texts for students with cognitive or reading challenges
- Study support: Tools can help students summarize notes, generate practice questions, or test their understanding before an exam
Assisting Teachers and School Staff
- Lesson planning and content creation: AI tools can help draft lesson outlines, worksheets, rubrics, and quizzes, reducing prep time and allowing teachers to focus on instruction
- Grading and feedback: Automated grading for short-answer or essay-style questions could free up time and offer consistent feedback
- Communication: Teachers could use AI to write emails to parents, generate newsletters, or translate communications into different languages for multilingual families
- Professional development: AI could assist teachers in learning new subjects or pedagogical techniques by offering curated resources or answering content-related questions on demand
Helping Administrators and Policymakers
- Performance analytics: AI could help identify patterns in student performance across schools or regions, flagging gaps in understanding or resources before they widen
- Curriculum improvement: By analyzing how students interact with material, AI could inform decisions about which parts of the curriculum are effective and which need adjustment
- Resource planning: Predictive models could assist with enrollment forecasting, staffing, and identifying where additional support is needed most
These use cases are not speculative. Most are technically feasible today. What they require is careful, responsible integration.
While AI has great potential in education, it is essential that implementation is done in a deliberate and thoughtful way. Many things can go wrong if it is rushed or poorly designed. For students, the risks are more visible and easy to empathize with. These include dependency, misinformation, privacy issues, and unequal access.
But it is equally important not to overlook the teacher's experience. Teaching is already a demanding and often thankless job. AI should not be introduced in ways that add complexity, increase pressure, or reduce autonomy. It should support and elevate the work of teachers, not undermine it.
If we want AI to play a positive role in education, we must ensure it works for both students and educators.
2. Learning About AI: Why Education and Literacy Matter
Students are already using AI, sometimes to deepen their learning, but far too often to cheat. That alone is reason enough to make sure AI is not just a tool they access, but a topic they understand. And not just students. Teachers also need support to teach and use AI meaningfully, ethically, and with confidence.
According to the AI Index, most countries are still behind. Only a few, including Ghana, South Korea, and the Netherlands, have explicitly embedded AI education in national curricula. But momentum is building, and several governments have recently taken more ambitious steps:
- United States: A 2025 executive order created a national task force on AI education. The plan calls for integrating AI into K-12 curricula, expanding teacher training, and building partnerships with industry to support implementation
- China: Starting in September 2025, AI education will be mandatory across all primary and secondary schools. The national curriculum includes AI fundamentals, programming, machine learning, and robotics
- United Arab Emirates: The UAE's "AI Skills" program aims to train one million people, including students and teachers. AI is already mandatory in some schools and is supported by national platforms and international partnerships
Despite this momentum, readiness remains a challenge. In the U.S., fewer than half of computer science teachers say they feel equipped to teach AI. In elementary schools, that drops to just 34 percent.
To move beyond slogans and announcements, we need clarity on what AI education actually means.
UNESCO's two global frameworks, one for students and one for teachers, offer a structured answer. They define the mindsets, skills, and knowledge learners and educators need in an AI-driven world.
Here is what that looks like in practice:
- Understanding the foundations: Students should learn what AI is and how it works, including concepts like data labeling, pattern recognition, and training models. Teachers need to go further by understanding not just how systems are trained, but how to choose, evaluate, and adapt AI tools for different learning contexts.
- Exploring real-world applications: Students should examine where AI appears in their daily lives, such as in recommendation systems, facial recognition, or virtual assistants, and question what those systems are doing and why. Teachers are expected to introduce these examples, link them to curriculum goals, and lead critical discussions about their impact.
- Ethical and critical thinking: Both students and teachers should be able to identify bias, understand how algorithms may reinforce inequality, and evaluate the transparency and fairness of AI systems. Teachers also have the role of guiding ethical debates, setting norms for responsible use, and modeling digital citizenship.
- Human-centered design and agency: Students are encouraged to design simple AI solutions that reflect principles of inclusion, accessibility, and sustainability. Teachers should support this process by helping learners make intentional design choices while ensuring AI remains a tool, not a decision-maker.
- Using AI for professional and academic growth: Students can use AI to explore topics, get feedback, or create projects, but they must understand the limitations and risks of doing so. Teachers are encouraged to use AI to support lesson planning, collaborate with peers, and stay up to date in their practice, while maintaining clear pedagogical judgment.
These competencies are not just technical. They are also civic, ethical, and human. They help build the capacity to question technology, shape it with purpose, and understand its place in our shared future.
3. AI Education Beyond the Classroom: For Everyone, Everywhere
The AI revolution is too big, too fast, and too far-reaching to be treated like a school subject that ends at graduation. Its impact will rival, and likely surpass, that of computers and the internet. In a world being reshaped by algorithms, models, and automation, learning about AI cannot be confined to classrooms or age brackets. Everyone needs it, and most people are not getting it.
In theory, AI is already transforming the workplace. It can automate reports, summarize research, accelerate analysis, and generate creative ideas. In practice, many professionals either do not realize what these tools can do or are using them just enough to sound like they do. The problem is rarely complexity. It is overconfidence, lack of time, or simply not knowing where to start.
From my own experience, you would be shocked by how little some professionals actually understand about AI. And I am not only talking about general users. I am talking about so-called 'AI experts' who fake it but never make it, who talk about AI for hours a day without ever really understanding how it works or what it can actually do. They use AI, technically, but not meaningfully. Some are building entire narratives about the technology without fully grasping what it is capable of today.
If AI literacy is going to matter, it cannot stay locked inside schools. It has to extend into companies, government agencies, creative industries, and public life. That means more than the occasional workshop or online course. It means real investment in training programs, accessible tools for self-learning, and a culture that expects people to understand the systems that are shaping their lives.
This is not just about productivity. It is about agency. AI is changing how we learn, work, make decisions, and interact with the world. And in a future defined by these systems, we cannot afford to let only a small group understand how they work while everyone else simply adapts.
Conclusion
AI is quietly rewriting how we learn, work, and make sense of the world. The question is no longer whether to teach it, but how fast we can catch up. If we want a future where people are not just using AI but shaping it, then education, in every form and at every stage, has to lead the way.
A 2025 meta-analysis of 51 experimental studies found that ChatGPT can have a large positive effect on students' learning performance, and moderate benefits for learning perception and higher-order thinking. These effects were strongest when ChatGPT was used in problem-based learning settings, when its role (such as tutor, partner, or tool) was clearly defined, and when applied over a sustained period of 4 to 8 weeks. In contrast, short-term or unstructured use often produced minimal or inconsistent results.
The message is clear: technology alone does not transform education. Its impact depends on how it is used, with purpose, appropriate context, and sufficient support. As more schools and governments adopt AI in classrooms, research like this should serve as a guide. We do not just need AI in education. We need it implemented thoughtfully and effectively.
