How will we give personalized education to 2 billion kids?
Humanity have hit a jackpot - while an estimated 2 billion kids need to be educated globally in the developing world, revolutionary new artificial intelligence has mastered skills thought to be exclusively human, like creating original text and art. Powerful systems like ChatGPT can already pass exams, fueling predictions that AI tutors tailored for personalized education will displace traditional classrooms. Yet risks abound: will teachers lose core skills if AI assumes too much work? Can complex emotional and social connections be coded? As debate rages, this article analyzes which human teacher qualities advanced AI could feasibly replicate today versus skills likely to elude algorithms. It finds widespread ethical, cognitive, and technical barriers persist before AI can fully transform learning. Can we ethically leverage AI to expand access and opportunity to vulnerable youth worldwide before more complex capabilities emerge? Grappling with this question may be the greatest challenge for the future of education.
ChatGPT and other AI systems that can generate fluent text, art and code have captured public fascination recently. Behind the hype, a serious debate simmers on if and how these technologies should be integrated in education. Some like Professor Daniel Andler warn deploying chatbots as “automatic talker-writers” risks “de-skilling” students by eroding foundational skills. Others such as Sir Anthony Seldon predict classrooms globally will have personalized AI tutors within the decade that adapt to individual learning needs and styles. Who's right? In this article I examine the key qualities of the best human teachers, assess which are within reach of current AI, and suggest where emerging capabilities still fall short. My aim is to spur deeper discussion on how to maximize benefits and mitigate risks as this extraordinarily disruptive technology will inevitably moves into the classroom.
"...Finch Weinberg is a middle-aged robotics engineer living in a post-apocalyptic USA. Finch knows that he is dying and he has built an AI robot to take care of his canine companion after he will be gone. The robot’s name is Jeff, and he learns to be good friend to Finch and caregiver to Goodyear, while they travel across the USA."
This is a 59-word AI summary of the beautiful movie “Finch” where Tom Hanks is playing the main character. The movie was released in 2018 that is year 4 B.C. (“Before ChatGPT”), and yet it captures the essence of generative AI much better than 85% of recent scholarly articles.
Jeff, who is an AI robot, is kind, caring, patient, emotional, emphatic, with a good sense of humor, highly adaptive, curious, and always learning. Robot Jeff appears more human than many people, solving complex tasks helping his human friend, making honest mistakes, learning from own mistakes and from Finch’s reactions, surprising Finch with new abilities and unorthodox ideas. One might suggest that at least 100 million individuals feel this way about ChatGPT.
There has never been a shortage of humanesque automatons in science fiction cinema and literature. But in reality, we tend to see technology, such as AI robots, only as an instrument, and we are not ready to accept a machine as a “virtual being” which possesses its own agency, intent, and cognition.
Few individuals on our planet have contributed to technological progress as profoundly and as early as Douglas Engelbart (1925-2013). Engelbart was one of the first innovators designing human-computer interaction in the form we see it now. His famous “Mother of All Demos” presentation at the Fall Joint Computer Conference in San Francisco on December 7, 1968, showcased pioneering technologies including the computer mouse, hypertext, and early video conferencing. Legend has it that at this very event, Engelbart famously proclaimed: “Technology should not aim to replace humans, rather amplify human capabilities”.
However, 55 years later in 2023 at the Digital Learning Week hosted by UNESCO, sentiments seem to have shifted. Renowned French mathematician and philosopher Professor Daniel Andler was referring to Generative AI systems as “automatic talker-writers”. In a thought-provoking conversation with prominent computer scientist Professor Stuart Russell, Andler suggested rapid expansion of ChatGPT adoption in the education process may not be prudent. Specifically, he warned: “If the teacher or co-teacher is not a real person but rather an automatic talker-writer, the learning process could be profoundly perturbed”.
Elaborating further, Professor Andler, professor emeritus at the Sorbonne since 2015, painted a vivid picture: “ChatGPT poses as a person whether outside the classroom when it acts as a private tutor for the student, or inside the classroom when deployed as a co-instructor alongside the teacher. This makes ChatGPT an autonomous instructor rather than a simple teaching aid or tool, which could have disruptive implications”.
Clearly Professor Andler’s skepticism towards embracing ChatGPT in academia stems from his perspective of ChatGPT as an independent instructor entity with some semblance of “personhood”. While Andler’s caution may appear excessive, he nonetheless assigns sufficient agency and humanity to ChatGPT to conceptualize it as a “virtual being”. This vision of anthropomorphized AI is likely not foreign in contemporary discourse concerning the imminent future of education.
In fact, esteemed British historian and education reform advocate Sir Anthony Seldon voiced comparable notions at the 2017 British Science Festival. Specifically, he predicted “AI robots will replace teachers within classrooms in under a decade”. Furthermore, in his highly influential 2018 book entitled “The Fourth Education Revolution: Will Artificial Intelligence Liberate or Infantilise Humanity”, Seldon elaborated his compelling vision for making personalized AI-powered education ubiquitously accessible.
“This is beyond anything that we've seen in the industrial revolution or since with any other new technology. These are adaptive machines that adapt to individuals. They will listen to the voices of the learners, read their faces and study them in the way gifted teachers study their students".
“Programmes will learn to read the brains and facial expressions of pupils, adapting the method of communication to what works best for their learning style. The personalized nature of the robots will enable pupils to learn new material at their own pace, rather than as part of a class. This is revolutionary for one-to-one learning”.
“It will open up the possibility of an Eton or Wellington-style education for all, everyone can have the very best teacher and it’s completely personalized; the software you’re working with will be with you throughout your education journey”.
We can assume that there are at least 2 billion school age children in the developing world, and if we want to provide each one with a personalized learning experience, we need to create 2 billion teachers.
A beautiful research paper “The AI Revolution in Education: Will AI Replace or Assist Teachers in Higher Education?” from University of Hong Kong researchers, Cecilia Ka Yuk Chan and Louisa H.Y. Tsi, presents a structured view at the challenges that AI will face when teaching students.
Emotional and Interpersonal Skills
1. Human connection: The emotional bond and interpersonal skills that teachers have are essential for students’ personal growth and development. /This will be very easy to implement. In order to create a bong connection with a student AI teacher just needs to study relevant information about her/him and change own behavior accordingly. This is very easy to implement by giving the system students’ context and introducing some Q&A mechanics into introduction session./
2. Cultural sensitivity: Teachers can understand and navigate the cultural nuances of their students and adapt their teaching approach accordingly. /This will be very easy to implement. AI teacher will develop cultural sensitivity in a same way as described above./
3. Developing resilience and perseverance: Teachers play a vital role in helping students develop resilience and perseverance by offering support, guidance, and encouragement in the face of challenges. AI systems may not be as adept at providing this emotional and motivational support. /This will be very easy to implement.AI teacher can provide emotional and motivational support in a same way as described above./
4. Building trust and rapport: Teachers build trust and rapport with their students through personal interactions, which helps create a positive learning environment. /This will be very easy to implement. With simple prompt engineering we can ensure that AI teacher is using appropriate wording and references to build trust and rapport with students./
5. Social and emotional learning: Teachers support students’ social and emotional learning by modeling appropriate behavior, discussing emotions, and helping students develop self-awareness and empathy. /This is already implemented in basic interaction with LLMs. We might need some simple tweaks in prompt engineering to enforce social and emotional learning by students./
6. Role model: Teachers serve as role models for their students, demonstrating a passion for learning, a commitment to personal growth, and the importance of hard work and perseverance. /This might be very difficult to implement, almost impossible, as AI do not possess history of actions. It is not likely that students will see AI teachers as role models. On the other hand, a strong argument can be made in favor of highlighting student’s own parents or community leaders as role models, and AI can help here./
Pedagogical Skills
1. Real-world context: Teachers can provide real-world examples and experiences, and this helps students better understand and relate to the material being taught. /This is already implemented in basic interaction with LLMs. There is no doubt that AI teacher will outperform most human teachers in preparing class material./
2. Encouraging curiosity: Teachers can inspire students to be curious by sharing their own enthusiasm for learning and fostering a growth mindset. /This will be very easy to implement. We certainly shall do prompt engineering that will keep AI teacher inspire students’ curiosity./
3. Professional development: Teachers continuously engage in professional development to improve their teaching practice and stay up-to-date with the latest educational research. /This is already implemented in basic interaction with LLMs./
4. Encouraging debate and open-mindedness: Teachers can foster an environment of open-mindedness and encourage debate by presenting diverse perspectives, asking challenging questions, and facilitating discussions. /This is already implemented in basic interaction with LLMs./
5. Conflict resolution: Teachers often help mediate conflicts between students and teach them essential conflict resolution skills. /This feature haven't been implemented but there are extensive papers explaining how to do this. AI teacher will need to intake additional information in order to help students with mediation./
6. Experiential learning: Teachers often design and facilitate hands-on, experiential learning opportunities for their students, such as field trips, lab work, or other interactive experiences. /This feature haven't been implemented but there are extensive papers explaining how to do this. AI teacher shall be supported by integration with other IT systems that will support learning experiments. AI teachers will perform this activity much better than humans./
Holistic Competency Development
1. Adaptability: Teachers can adjust their teaching methods and strategies based on the specific needs of their students. /This will be very easy to implement. As soon as AI teacher will be equipped with feedback collection (in a form of dialog, or otherwise) it will excel in adaptability./
2. Critical thinking and creativity: Teachers can foster creativity and critical thinking in students by designing engaging lessons, asking thought-provoking questions, and encouraging open discussions. /This is already implemented in basic interaction with LLMs./
3. Collaboration and teamwork: Teachers help students develop collaboration and teamwork skills through group projects, discussions, and other cooperative activities. /This will be very easy to implement. AI teacher has to access students’ interaction to measure and guide their collaborative work./
4. Nurturing creativity and innovation: Teachers help cultivate creativity and innovation by allowing students to explore new ideas, take risks, and think outside the box. /This is already implemented in basic interaction with LLMs./
5. Teaching life skills: Teachers often help students develop essential life skills, such as time management, goal setting, and decision making. /This is already implemented in basic interaction with LLMs./
Personalized Support
1. Behavior management: Teachers are skilled at managing classroom behavior and addressing issues as they arise. /This will be very easy to implement. Designers of AI teaching process have to AI teacher has to access students’ interaction to measure and guide their collaborative work./
2. Support for students with special needs: Teachers provide tailored support to students with special needs, accommodating their learning styles and addressing any unique challenges they may face. /This will be very easy to implement. We should select and provide AI teacher with instructional material that explain the best ways to interact with students with special needs./
Community and Civic Engagement
1. Parent-teacher communication: Teachers communicate with parents to discuss students’ progress, share concerns, and offer guidance. /This feature haven't been implemented but there are extensive papers explaining how to do this. A challenging task of building rapport between AI teacher and student’s family has to accommodate cultural differences between various cultures./
2. Encouraging civic engagement: Teachers help students understand the importance of civic engagement and develop a sense of responsibility to their community and society. /This is already implemented in basic interaction with LLMs./
Career and Personal Mentorship
1. Career guidance and mentorship: Teachers can offer valuable career guidance and mentorship by sharing their own experiences, offering advice, and connecting students with relevant resources and opportunities. /This is already implemented in basic interaction with LLMs./
Physical and Artistic Education
1. Physical education and sports coaching: Teachers play a crucial role in promoting physical fitness and coaching sports teams. /This will be very easy to implement. AI teacher will focus on developing students’ collaboration in physical exercises, along with measuring and analyzing performance./
2. Artistic expression and appreciation: Teachers help students develop their artistic skills and appreciation for various art forms. /This task might be one of the most challenging to AI teacher, as it involves advanced technical tools needed to capture students’ activity and substantial computing power to analyze comprehensive data on artistic expression. But there are at least 100,000 teams who know how to do this./
Most human teacher skills and qualities could be replicated in an AI teaching assistant without too much effort. However, complex social and emotional capabilities would remain challenging.
Cecilia Ka Yuk Chan and Louisa H.Y. Tsi rightly stress ethical issues on using AI in education, like privacy, bias, and content misuse. These issues deserve focus. But I want to note AI developers do care about ethics too. Companies, professional association, government agencies, and multilateral conferences also devote attention to these topics. I believe solid solutions will emerge soon.
Sir Anthony Seldon have predicted that AI teachers can be placed in classrooms by 2027, personalizing education for each student. Our top priority is to provide AI teachers for the estimated 2 billion kids in developing countries who need good teachers the most. Let's get moving on making this helpful AI teaching tool come true for that big group of children as soon as we possibly can.
Let AI elevate learning worldwide!