AI Is Rewriting High School and College Education

Among the many questions that are fueling the debate on Artificial Intelligence (AI), some of the most fascinating concern the education system. Just as digital has helped redefine parts of secondary and tertiary education curricula over the last fifteen years, ChatGPT and its recent derivations call for a major rethinking that will significantly impact the life of schools and universities around the world.

First of all, to know how to use the tools that AI makes available to us, it is essential to have at least a minimal knowledge of coding, what goes under the rubric of Computer Science. This in turn depends on having a good grounding in Mathematics and also Logic – which is often complemented by foundations of Critical Thinking. Pedagogy teaches us that these disciplines should be taught from an early age and it is therefore essential that primary and secondary schools equip themselves to help the newly spawned Beta generation (those born from 2025 onwards) to face AI moving downslope rather than uphill, as is instead conversely the case for all the generations that preceded it.

But always with reference to skills, this requires an effort on the part of universities to integrate training in Computer Science skills in their degree programs, regardless of the specific disciplinary fields chosen by students. Bocconi, although it is a university specializing in the field of Social Sciences, has already done so for some years now, and it is nice to see the tendency of national and international universities going in this direction, which means both strengthening STEM departments and integrating them into portions of their educational offer in an innovative way.

Moving along the path linked to the development of skills, we must not be fooled by the idea that AI tools will progressively become simpler with easy-to-use interfaces, as ChatGPT already highlights, as our ability to manage the tool will depend on the knowledge of its design and functioning, as well as on the use of the data it produces.

But regardless of the skills, the challenge launched by AI focuses on the method. In fact, the rapid diffusion of ChatGPT, Dall-E, Gemini, Claude, DeepSeek and all the tools that are based on Machine and Deep Learning models have consequences for the two fundamental spheres of higher education: research and teaching.

In fact, Large Language Models and Computer Vision provide researchers around the world with data analytics tools hitherto unthinkable. This is confirmed by some extraordinary discoveries that are taking place in the fields of health, energy and security. A key example of this is the recent Nobel Prize in Chemistry awarded for solving the protein-folding problems with the help of the AI models developed by Google's DeepMind.

These extraordinary opportunities are counterbalanced by non-trivial risks of artificial data production and plagiarism, which are pushing scientific journals to adopt rigorous ethical rules and strong gatekeepers. Just as is happening in the world of media an entertainment, whose business models are also based on the production and circulation of information and knowledge, it will be important to work on and consolidate the rules of engagement to avoid cases of counterfeiting and plagiarism that compromise the quality of the services offered and the knowledge produced.

But then there is teaching, the basis of the revenue model of any university. The provocative article by Chris Terwiesch of the University of Pennsylvania has shown that ChatGPT is capable of passing even complex MBA exams, and, if well trained, it could also help write exams shortly. In short, it turns into potential Teaching Assistant, and professors from various corners of the world have already been employing it this way for a few semesters. In their beautiful book on the use of AI for decisions, Karim Lakhani of Harvard together with Marco Iansiti posit the use of AI to build chatbots that act as real virtual assistants for students, further raising the level of learning. Thinking of boomers like the present writer, it is funny to reflect on the fact that not long ago people were going to class to decode the content of books written cryptically. On the contrary, Generation Z and Generation Alpha have had the opportunity to study from interactive textbooks with videos and links that have allowed a substantial enhancement of the classroom experience. It is no coincidence that in the last fifteen years teaching has evolved substantially compared to the last century, not only by proposing forms of online education and hybrid learning, but above all by modifying the interaction in the classroom itself that can take part of the learning path for granted. 

What will plausibly happen soon is that Generation Beta will have the opportunity to study with chatbots of the type we use today to book a flight or question a bill - let's just hope they are more empathetic. This will dramatically increase the possibility of providing more personalized learning and will lead to a substantial rethinking of classroom teaching hours, given that the student will be able to independently learn part of the baggage of the exam of the specific discipline by interacting with a virtual assistant. The future of teaching is therefore destined to reduce the time spent in the classroom and to amplify the time spent receiving students - the so-called office hours - and experimentation to further enhance the learning path of future students.

In short, we are living through a truly fascinating moment in history that will lead to major changes that will further improve the educational system. The theory of innovation, which is rich from the empirical evidence matured over the last hundred years, teaches us that the speed of radical innovation is much quicker than the pace of institutions, which often try to slow it down and shape it according to their own purposes. The institutions in question, School and University, are so crucial for society, but also resilient and not very inclined to change. Since in this case we are talking about truly enhancing knowledge and learning, i.e. the kind of bread Sapiens feeds on to evolve, we hope the utter relevance of AI may be fully understood, so that we try to govern its impact without trying to arrest its development.