We’re Asking the Wrong Question About A.I. in Education

If you traveled back in time to the end of the 16th century and visited the Laurentian Library in Florence, you would discover that the books—precious manuscripts from classical civilization—were chained to the desks. Looking back five centuries makes it easy to see how far we’ve come, and how technological and societal advancements have dramatically widened access to information and learning. 

Knowledge and learning were once rare and expensive. Today, they are ubiquitous and cheap. Instead of traveling to Florence to consult a book bolted to a reading desk, you can access an online version, join remote discussions, sign up for online university courses from your living room or listen to experts share their views on podcasts during your commute. 

From the invention of the alphabet to the rise of the internet, information technology revolutions have democratized learning. They have also repeatedly transformed the role of teachers, as educators adapted to the new tools at their disposal. The advent of generative A.I. will undoubtedly reshape the relationship between student and teacher once again. 

Generative A.I. is already playing a crucial and growing role in education. Tools like Khan Academy’s Khanmigo and Google’s A.I.-assisted classroom features are actively in use in schools around the world. The OECD and UNESCO have both issued frameworks cautioning against over-automating the teaching relationship. The technology is here, it is scaling rapidly and the decisions we make about how to deploy it are being made right now, making the question of what A.I. should—and should not—do in education more pressing than ever. 

As someone working at the forefront of online tutoring, where technology and human education come face-to-face, a question I get asked is: “When will A.I. teachers fully replace humans?”

The answer is simple: never. 

As we stand on the cusp of this next information technology revolution, we should instead ask a more productive question: what role should A.I. fulfill in the learning process, and, more crucially, how do we understand the role of teachers in a world where knowledge has never been easier to access?

One way to approach this question is by asking what A.I. can actually do in education, and the capabilities are hugely impressive. Generative A.I. has first-class transcription, synthesis, pattern-matching and content-generation capabilities. Today, it is already possible to generate a language student’s class transcript, use it to identify and summarize mistakes, build a tailored lesson plan and generate personalized homework for them. It can even remind the student to complete that homework, mark their responses and brief their teacher on what to prioritize in the next class. 

It is now also  possible to generate live practice sessions with human-like avatars that respond in real-time and correct your mistakes. These capabilities offer real value for self-directed learning and student assessment. The nature of the interaction itself can be an advantage: an A.I. “teacher” will not judge you. It will provide a low-stakes, comfortable environment to learn, with no social risk or pressure. 

However, simply asking what the technology can do misses a more important question: what are the unique educational powers of human teachers? 

Books did not replace teachers. Neither did apps nor online videos. MIT offers thousands of hours of learning materials for free, but the university continues to charge substantial fees for a human-based education. This is because human teachers have unique abilities that play a more important role in education than the content generation and assessment at which A.I. is effective. 

Teachers motivate, inspire and make judgment calls in ambiguous situations. An A.I. tool can tell you your mistakes, but a teacher can give you the motivation to try again. An A.I. app can show you the next step in your learning journey, but a human teacher can help you remember why you are learning in the first place. While A.I. can create a comfortable environment for learning, a teacher can judge when you need to be pushed beyond your comfort zone. Most critically of all, the relationship between teacher and student forms a social contract of shared commitment and accountability. Apologizing to ChatGPT for not doing your homework just doesn’t carry the same weight. 

There is a prevailing A.I. thesis today that is essentially nihilistic: humans are replaced and costs are saved. Taken to its conclusion, this leads to an education system that is less human and, paradoxically, less effective. But it doesn’t have to be this way. We can set a more aspirational goal: to use the power of generative A.I. to bring more humanity to education. By leveraging the complementary strengths of both human teachers and intelligent tools, we face an opportunity to define a future of education where more A.I. means more humanity. 

To understand what that future could look like, we must start with the present reality of how educators are actually spending their time. In the U.K., around 50 percent of school teachers’ working hours are consumed by administrative tasks, including lesson preparation, writing pupil guidance, setting assessments and general management tasks. Teachers consistently highlight the high preparation and communication burden of their role as a primary source of stress, compounded by disconnected systems and over-stretched administrative staff. The OECD’s most recent TALIS survey found that teachers across member countries consistently report working beyond their contracted hours, with administrative burden cited as a leading driver of burnout and attrition.

This isn’t unique to the school environment. Individual self-employed tutors on Preply typically spend over 30 minutes doing lesson preparation and admin for each hour of time spent teaching. Teachers and tutors consistently state that time with students in the classroom gives them the most satisfaction, but for many, the proportion of time they spend with students is far too low.

Student assessment, lesson preparation and general administrative tasks are precisely where A.I. excels. Redirecting more of these tasks to A.I. frees teachers to spend their time on the unique educational strengths which only humans can provide: motivation, inspiration, social accountability and calibrating the level of challenge offered to each student.

This is why asking “when will A.I. teachers replace humans?” is the wrong way to think about the role of A.I. in education. Instead, by asking how we use A.I. to bring more humanity to education, we can be more intentional about designing an education system that combines the best of human teaching and technology.

Education is too important to be focused on cost-saving or an outsourcing opportunity. We should be ambitious. The goal should be a future where every learner has access to a great teacher, where the relationship with the teacher is personal, the curriculum is tailored to their needs and interests and everyone is supported in pursuit of excellence.