Those who talk about the impact of AI on the future of work tend to fall into two broad camps: A) those who think there will be mass surplus labour (and therefore unemployment) in the future, and B) those who are confident that jobs that are lost will be replaced, or more than replaced, by new jobs. No matter what the net impact of AI is on employment, we think we will need major transformations in the education system to get us ready for a future with more AI. Around the world, many schools are preparing for the future by spending more classroom hours on science and mathematics, and teaching kids to code. This is important, but we think that helping students to develop their human specialties is more important still.

‘Human specialties’ is a term we use at – Dr. Rami Shaheen to describe those skills at which people outperform machines. It happens that many of these skills are also those that seem particularly close to our humanity, like emotional intelligence, creativity, and collaboration. A modern curriculum needs to incorporate both human specialties and STEM (Science, Technology, Engineering and Mathematics) subjects. We don’t think this is a question of what subjects we get rid of, but rather how we teach students.

Our kids will need to learn less like automatons for them to thrive in a world of automation.

While what exactly constitutes ‘human specialties’ remains contested, a research paper produced by the New South Wales (NSW) Department of Education in Australia gives a good overview of current thinking. The NSW Department of Education describes ‘21st century skills’ as encompassing nine areas: critical thinking, creativity, metacognition, problem solving, collaboration, motivation self-efficacy, conscientiousness, and grit or perseverance (see page 19 of the NSW research paper, ‘Key Skills for the 21st Century’, for a closer examination).

There has been extensive research both about how to teach these human specialties, and why doing so will be particularly important in the age of AI. But there is not much guidance combining the ‘how to’ with the ‘why’ in order to provide sensible policy recommendations. Notable exceptions we’ve found are work by the NSW Education Department and the Organisation for Economic Cooperation and Development (OECD) project, OECD Education 2030.

So when it comes to teaching human specialties there are many outstanding questions and challenges that need to be addressed.

First, how we can best teach human specialties? A paper produced by the Education Department of NSW gives examples of countries that have actively implemented some or all of these skills into their curriculums: such as Finland, Australia, areas of Canada and the United States. The policies that seem to have worked best in all cases include promoting the adoption of these skills into school curriculums at all levels (primary and secondary), encouraging collaborative projects within all subjects and allowing teachers to have the freedom to encourage these skills and asses them as they see fit. Making these kinds of changes means more than changing the curriculum: it’s also about giving more scope for teachers to find creative ways to teach skills, more opportunities for professional development for teachers, as well as rethinking the current teaching and learning environments.

Second, if schools are focusing on different skills, how should they be assessed? This is an important question, because school curriculums will end up looking very different depending on the answer. Part of the OECD’s Education 2030 project is to establish a framework of assessment for human specialties, which they hope to launch in 2019.

Does preparing our kids mean continuing to test them more, or should we start to redesign our curricula so they don’t rely so heavily on examinations? Richard Watson, in Chapter 3 of Future Frontiers: Education for an AI world, asks us to rethink how we assess students. He gives a number of suggestions, including ending our ‘obsession with exam results and league tables’ and getting rid of exams altogether or finding new ways to assess human specialties. Watson points out that part of why this is such an important question is that we are moving (even further) into a knowledge-based economy. We need children to innovate, think for themselves and work together. The current system of individual, standardised tests does not encourage this. So, if we really want to test human specialties, we need to think of a way to do so that is not discordant with what we are teaching.

Third, how can we enable a curriculum to be more adaptable and take into account new research, as and when it comes out? The United Nations Sustainable Development Goal number 4 (SDG4) gives a number of targets to reach by 2030 including ‘complete free, equitable and quality primary and secondary education leading to relevant and effective learning outcomes’ and ‘substantially increase the number of youth and adults who have relevant skills… for employment’. In order for countries to achieve these goals by 2030, work needs to start quickly. The problem of teaching ‘relevant skills … for employment’ is that they are taught years before students enter the labour market. Thus the task is not to teach the skills relevant for today’s labour market, but to sensibly predict which skills will be needed in many decades’ time: which is why we urge a greater focus on the development of human specialties.

The idea of adding these kinds of skills to school curriculums is not a new one, but honing human specialities has never been more important than it is now. This means moving quickly and flexibly and trusting the experts, teachers. Teachers need space and time to invest in their own understanding so they can also give students the broader education they need. We will need many more STEM experts in the decades ahead – but also empathetic, creative and thoughtful citizens who know which problems can be solved by machines, and which need a human.