By Dirk Van Damme
Head of the OECD Centre for Educational Research and Innovation
Our world and our way of living are going through a huge shock. With the exception of the oil crisis of the early 1970s and the collapse of the financial system in 2008, we have experienced 75 years of fairly unbroken growth and social progress. This, now, has suddenly come to an end. The pandemic heralds a new era in which the world must live with recurring systemic shocks. Experts tell us that this pandemic will not be the last one. And the full impact of climate change is yet to come. Still, enormous as these shocks will be, it may be something entirely else that will trigger the deepest and most disruptive change in the 21st century: artificial intelligence (AI).
Still in the early days of its expansion, but AI’s impact on so many different spheres of life is already obvious. Artificial intelligence is the result of scientific and technological progress in the fields of computing and mathematics, driven by big data and machine learning, but also fed by insights in brain research, neuroscience and cognitive psychology. Quantum computing will soon dramatically move the frontiers of computer capabilities. And integration with synthetic biology is just around the corner.
Its applications have penetrated the world of production and labour through robotics, but also seemingly innocent interventions in consumer behaviour, leisure preferences and cultural habits, more contested domains of surveillance or political manipulation, and even education. Artificial intelligence is turning our historic towns into ‘smart cities’ and our classrooms into ‘smart learning platforms’. Indeed, we are starting to discover the promises of AI-driven learning management tools and adaptive courseware. And many more educational applications are still to come that will help make teaching and learning more effective and more equitable.
And yet, there is a great deal of concern about AI. The initial worry that artificial intelligence would lead to massive job losses seems to gradually fade away. Concerns are now focusing on the risks of AI being used unethically. In May 2019, the OECD set out its Principles on Artificial Intelligence. The principles have been endorsed by a large number of countries and provide an ethical and political framework for policy development in this field. And for good reason, because derailments are a real risk.
As AI continues to become part of how we live as workers, comsumers, citizen and learners, questions about the balance between human and AI informed decision-making become important. More broadly, how should we see the relationship between humans and smart machines? This is not a trivial question, but a question with very deep implications for education. How will the capabilities of smart machines develop compared to what humans are capable of? Understanding the balance between computer capabilities and human skills will help education systems decide what students should learn at school. This is a priority research topic for the OECD Centre for Educational Research and Innovation. This project assesses AI capabilities, using tests available from education, occupational certification, cognitive psychology, and animal cognition, and compares these capabilities with what humans can do.
It makes no sense to design education systems that educate learners to do what computers and robots can do infinitely better
The implications for future curriculum development are possibly huge. It makes no sense to design education systems that educate learners to do what computers and robots can do infinitely better. What students need to learn at school will progressively be informed by what makes humans truly human. Activities such as ethical decision-making, the passing of legal judgments by judges, the diagnosis of a medical doctor, the esthetical appreciation of an artist, or the pedagogical relationship between the teacher and the student, they can be informed by algorithms and machine learning, but the human element will remain, especially when information is incomplete or contradictory.
It will still make sense to read Shakespeare, to probe the beauty of prime numbers or to look at a drop of water under a microscope. Human ingenuity is not a magical trick that works in a vacuum, but results from the interaction between complex cognitive processing of prior knowledge, well-developed character qualities and sophisticated values. AI will help humans to overcome the barriers of information-processing, so that human cognitive and non-cognitive processes can become stronger, and the final human value added becomes more sophisticated and effective.
Yet, there are important dimensions of human activity, which have somehow been snowed under in contemporary schooling and which should get more attention in an AI-enhanced future. Most obvious is the social-emotional domain. The social-emotional skills of an experienced nurse for example, so important for the recovery and well-being of patients in hospital, are very difficult to train even the most humanoid robots for. Robots are getting their place in hospitals and can significantly relieve the work of doctors and nurses. The daily preparation of a patient’s medication can easily be done by a robot, but the friendly and encouraging chat by the nurse definitely not.
The same is true in education itself. When schools closed and students had to turn to distance education and home schooling, and the physical and emotional presence of the teacher disappeared, students really suffered. The combination of cognitive incentivising, substantive rigour, emotional support and empathy, and caring, qualities which define a good teacher, is essential in fostering an effective learning environment. Students can do without, and the most experienced and capable of them will learn a lot from sophisticated, AI-supported educational resources, but especially more vulnerable or struggling students will not succeed if they have to miss the human factor of a real teacher.
The most experienced and capable students will learn a lot from sophisticated, AI-supported educational resources, but more vulnerable or struggling students will not succeed if they have to miss the human factor of a real teacher
There are several other domains beyond the social and emotional, which will need more space in curriculum development as AI is decreasing the load of repetitive routine cognitive tasks. We can think of the ethical and the esthetical, specific domains of competence and expertise which require adequate training and experience to blossom. We could also think of manual dexterity as an element of craftmanship. Why are highly sophisticated and valuable equipment still being advertised as manually assembled as a indication of quality? Thinking beyond today’s curricula into another world of possibilities for the kind of learning that makes humans flourish, is what AI invites us to do.
We can witness the many benefits of AI today. The development of vaccines against COVID-19, one of today’s most remarkable scientific and technological achievements, would not have been possible without smart machines to decode the genomic sequences of the virus and the artificial intelligence embedded in the technologies to crunch the data. Likewise, artificial intelligence will complement and enhance human ingenuity in finding solutions to overcome the consequences of climate change and to provide the opportunity to a good life for as many people as possible on this planet. In order to make these promises happen, education today needs to come to grips with artificial intelligence and prepare students for a different world than the one we inherited from the 20th century.
- OECD International Conference on Artificial Intelligence in Work, Innovation, Productivity and Skills
- OECD Centre for Educational Research and Innovation
- Child safety in the digital age: How education systems can help
- Coronavirus took school on line, but how digital is the future of education?
- Future proof? Four scenarios for the future of schooling