FIR0020

Written evidence submitted by University College London: Institute of Education (IoE), Computer science and the Institute of Health Informatics.

1. Evidence submitted by University College London: Institute of Education (IoE), Computer science and the Institute of Health Informatics. UCL. UCL IoE has an internationally leading reputation for excellence in education and research. We produce impact and innovations to enhance people’s lives globally. In this submission we focus in particular upon the Artificial Intelligence element of the 4th Industrial Revolution, whether interfaced through robotics or the Internet of Things, AI is the technology that can provide the intelligent platform to power the Fourth Industrial Revolution.

Executive Summary

2. We welcome this inquiry and in particular we applaud the chair of the Education Select Committee, Robert Halfon’s call for fast action to make sure that our education system and curriculum are ready to ensure that the UK benefits from the development of new technology, and his recognition of the particular needs of those from disadvantaged backgrounds. We are concerned that the UK is currently slipping further behind in terms of preparing our young people for a world of work and leisure that is being transformed by AI, robotics, automation and the Internet of Things (IoT).

3. Andreas Schleicher, head coordinator of the OECD Programme for International Student Assessment (PISA), has said that the UK is an area of concern because our curriculum is too narrow and we do too much rote learning, so we’re not preparing students effectively[1]. And in the education ‘pillar’ of the recent Economist automation readiness index we finished 20th out of 25 countries in preparing children for 21st century knowledge and skills[2]. That’s appalling. We’ve got to a point where we have to do something. The countries that make sure their populations are educated in this way will achieve the most when it comes to the 4th Industrial Revolution.

4. Recent government attention and investment has focused on science and technology education, which whilst important, is not the answer in and of itself to the challenge of the 4th Industrial Revolution. The answer is to be found in a radical reformulation of our education system founded upon an intelligence-based curriculum that concentrates on our Human intelligence beyond, as well as including, the particular form of intelligence that technology can reproduce and automate. What could be more important to any nation than the intellectual capacity of its population? We ignore this priority at our peril and yet that is currently what we are at serious risk of doing. If we continue to educate and assess our students on the knowledge and skills that can now be achieved by AI, we doom our population to a future of machine dominance and economic irrelevance.

5. AI is taking over a great deal of what has previously been viewed as the human domain. As a result, the evidence that we need to change the way we view intelligence and the way we design our education systems is increasingly compelling. We need to act on this evidence and use our human ingenuity to re-imagine our education systems to enable us to remain the smartest intelligence on the planet. The need for educators to be appropriately trained and equipped is critical. Educators need to work hand in hand with technology developers and research experts so that educators gain a voice in the AI technologies that will become part of their, and their students’ daily life.
6. We hope that this inquiry signifies a recognition of the need for profound and far-reaching change to our education system and look forward to assisting in whatever way that we can.

The interaction between the Government’s industrial, skills and digital strategies


1. The Prime Minister's speech on 21 May 2018 heralded the government’s investment into science and research and set out a vision for the future, with Britain leading the world in the 4th industrial revolution by nurturing the talent of tomorrow through: “an education system that gives young people the skills they need to contribute to the economy of the future.” This system will be grounded in a curriculum that sets the highest standards, proper support for teachers, more rigorous science GCSEs, more young people going on to do sciences at A-level, tax-free bursaries for teachers in priority subjects, new high-quality T-levels, Institutes of Technology, and a national re-training scheme. It is wonderful to see this recognition of the importance of STEM education and teacher education and training. However, as with the recently announced strategies upon which this speech was built, the focus on STEM alone misses the importance of a broad interdisciplinary education that also embraces the arts and humanities, creativity and imagination: these too are crucial if the UK is to lead the 4th Industrial Revolution. Education beyond computing and science and outside the University sector has been largely ignored, and this will be our downfall. We must start with our youngest learners, rather than repairing the damage at a later stage, if we are to benefit from what this revolution could really provide for society.

2. The Government’s digital strategy (https://www.gov.uk/government/publications/uk-digital-strategy/), launched in March 2017, promised a focus on skills through the creation of industry collaborations, international tech hubs, and digital skills partnerships, for example. The pledge was “to grow the UK’s technology skills” and to make it “a stronger, fairer country that works for everyone, not just the privileged few.” However, beyond the promise of coding in the National Curriculum from Key Stage One onwards, investment in a Network of Teaching Excellence in Computer Science and helping teachers and school leaders better understand technology, improved digital infrastructure for schools and the provision for computer science students of “the real-world, up to date skills needed in the digital economy,” there is little attention to the UK’s wider education provision. These technical skills cannot be successfully achieved in a vacuum and vital components for the future such as data literacy and ethics need a different approach.

3. Life sciences and AI are singled out in industrial, skills and digital strategies. Traditionally digital skills have been considered as a specialist domain and not integrated across all disciplines- that is clearly changing but needs to change even more. There is investment in maths, digital and technical education to address shortage in STEM skills but there also needs to be investment in the social sciences associated with digital services and AI, e.g. the ethics of data sharing and public response to AI. AI is dependent on the quality and the representativeness of the training dataset. If unrepresentative or inadequate data are not available, then the possibilities of AI being used are reduced, the impact of AI will be diminished and may even be harmful. Therefore, although there is mention in the life sciences digital strategy of digital innovation hubs negotiating better use of local data, this is not enough. In the NHS, fragmentation of services and data providers and data guardians has led to very poor interoperability of systems and data. Harmonisation ad interoperability of data are the bedrock of providing the right and the best data for AI and must be prioritised.

4. The government’s industrial strategy (https://www.gov.uk/government/topical-events/the-uks-industrial-strategy) pledged to raise total research and development (R&D) investment to 2.4 per cent of GDP by 2027 and identified AI and Data Economy as one of the Four Grand Challenges that the UK faces if it is to be at the forefront of the industries of the future. Key policies include establishing a technical education system “that rivals the best in the world” along with an additional £406m investment in maths, digital and technical education, and a new National Retraining Scheme to support people to re-skill. However, the Industrial Strategy Challenge Fund for research and innovation, published in May 2017, largely ignores education. Once again, the focus on the technology misses the point that to benefit from what the technology can bring we need people with imagination and creativity, a sophisticated personal epistemology and highly tuned metacognitive and emotional intelligence. One only has to look at the hiring done by tech companies to see evidence of the need for knowledge and skills beyond STEM .[3]

5. AI is about the interdisciplinary study of Intelligence, involving multiple disciplines including: psychology to help us understand human abilities such as problem solving, memory, vision and learning; philosophy of mind to shed light on what it is to be human, what consciousness is and why it is important for of our own intentional behaviour and for understanding the behavior of others; computer science, mathematics, and logic to enable us to build complex technologies that can process information at speed; and linguistics to explain the structure and functions of languages for communicating and thinking. The integration of so many different subject areas into the ‘discipline’ of AI, plus the breadth of activity that human intelligence affords have been somewhat overlooked of late as attention has focused upon machine deep-learning AI. They are however nevertheless fundamental to understanding Intelligence and developing AI.

6. The recent House of Lords Select Committee report on AI stated 5 principles, the fourth of which was that: “All citizens should have the right to be educated to enable them to flourish mentally, emotionally and economically alongside artificial intelligence.” We agree and we know that this cannot be achieved without attention way beyond science and technology education.[4]

7. Finally, we are concerned that Education is a poor relative in comparison to other government priorities – when it should be at the ‘top of the pile’. For example, the Topol Review: Preparing the healthcare workforce to deliver the digital future, whose interim report is published this month was commissioned by the Secretary of State for Health and Social Care, Jeremy Hunt, as a key element of the NHS workforce development strategy to ensure that it is at the forefront of life-saving or life-changing care for decades to come. The Independent Technology Review considers: How technological and other developments (including AI and Robotics) are likely to change the roles and functions of staff; the implications of these changes for the skills required and what this means for the selection, curricula, education, training, development and lifelong learning of current and future NHS staff. We need such a review for Education and we need it urgently.

The suitability of the current curriculum to prepare young people for the Fourth Industrial Revolution


8.     Few, if any, of the potential benefits of the 4th Industrial Revolution, from personalized medicine to increased productivity through automation, will be achieved at scale unless we address the educational and training implications of AI now.

Figure 1: The AI and Education Knowledge Tree (Adapted from[1])

9. The nature of what needs to be done is illustrated in Figure 1. There are two key questions to be addressed:

1. How can we use AI to improve education and help us address some of the big challenges we face?
2. How can we educate people about AI, so that they can use and benefit from AI and so that they can actively contribute to the debates and developments within AI in an informed way? Implicit in this question is the urgent requirement to educate the educators and trainers who will be expected to educate others.

10. There are three key parts that need to be introduced into the curriculum at different stages of education from early years through to adult education and beyond if we are to prepare people to gain the greatest benefit from AI. The first is that everyone needs to understand enough about AI to be able to work with AI systems effectively. This part is essential for AI and human intelligence (HI) to augment each other and for us to benefit from a symbiotic relationship between the two. For example, people need to understand that AI is as much about the specification of a particular problem and the careful design of a solution as it is about the selection of particular AI methods and technologies to use as part of that problem’s solution.

11. The second key part of the AI curriculum is that everyone needs to be involved in a discussion about what AI should and should not be designed to do. Some people need to be trained to tackle the ethics of AI in depth and help decision makers to make appropriate decisions about how AI is going to impact on the world. If we ignore the need for education about AI, then we risk failing to empower people to make key decisions about what it should and should not, could and could not, will and will not, be able to do for society. We need AI literacy in the same way that we need Scientific Literacy. Just as we would expect modern citizens to have some basic science knowledge to be able to operate and effectively contribute to the decision-making process of a society, we should expect people to have basic knowledge of AI to be able to effectively operate in an AI world. We would also expect them to contribute to decisions on what an AI should or shouldn’t do which would require them to have some kind of AI literacy.

12. The third part of the AI curriculum is that some people also need to know enough about AI to build the next generation of AI systems. If teachers are to prepare young people for the new world of work, and if teachers are to prime and excite young people to engage with careers designing and building our future AI ecosystems, then someone must train the teachers and trainers and prepare them for their future workplace and its students’ needs. This is a role for policy makers, in collaboration with the organisations who govern and manage the different teacher development systems and training protocols. The need for young people to be equipped with knowledge about AI is urgent, and therefore the need for educators to be similarly equipped is critical and imperative.

13. On a more positive note, the development of AI teaching assistants will provide an opportunity for developing deeper teaching skills and enriching the teaching profession. This deepening of teacher expertise might be at the subject knowledge level, or it could be concerned with developing the requisite skills to support and nurture collaborative problem solving in our students. It could also result in teachers developing the data science and learning science skills that enable them to gain greater insights from the increasingly available array of data about students’ learning. Any failure to recognise and address the urgent and critical teaching and training requirements precipitated by the advancement and growth of AI is likely to result in a failure to galvanise the prosperity that should accompany the 4th Industrial Revolution.

14. An education system that focusses on developing the routine cognitive skills that can most easily be automated makes little sense in today’s world. A much more sensible and strategically valuable approach would be to focus on using teaching and training approaches that develop metacognition and self-efficacy: two concepts that are inter-linked and essential for lifelong learning. We also need to increase people’s understanding across and between disciplines and their ability to work with others, both Human and Automated to solve real world problems.

15. A more sophisticated education system would ensure that our students are even smarter than ever before. It is much harder to develop a sophisticated personal epistemology that enables one to construct an evidence-based understanding of a complex and contested subject, than to learn and memorize features of that subject for future communication and application. It is often much easier to solve a problem alone, if that problem is tractable enough for one person to tackle, than to work with others to solve a more unwieldy and complicated problem. The intelligence-based approach is not an easy option. However, it does provide an opportunity for subjects, such as art and drama to find their way back into the curriculum with greater force. The intelligence-based approach depends far less on absorbing large amounts of academic information into memory, and far more on understanding how to construct understanding and commit knowledge to memory, when it is appropriate to do this, why it is appropriate to do this and for what purpose. 

16. The emergence of Intelligent networked devices and software provides a tremendous opportunity to increase our focus on neglected areas: Creativity and imagination, for example can be assisted through large bodies of knowledge that have been securely committed to memory. Creativity and imagination enable us to express our thoughts, feelings and desires and they underpin scientific and technological development too. Creativity and imagination can be nurtured by education, although systems that focus primarily on knowledge acquisition where there is an emphasis on testing and examinations can hamper learners’ capacity to be imaginative and creative.

17. A greater opportunity for Art and Drama could also be possible. Recent decades have seen increasing amounts of content in many school college and university curricula. Concerns have been raised about the extent to which subjects such as art and drama are being squeezed out by cost-cutting and/or in order to make way for other academic subject areas (see for example, Johnes, 2017; www.robinalexander.org.uk/wp-content/.../Alexander_Curious_Minds_July17.pdf). Re-thinking how we treat knowledge in the curriculum could make way for more time for subjects such as art and drama.

18. Some of the educational changes that are now essential will be easier to deliver than others. For example, we know that humans can excel at social interaction and that their abilities can be developed through education and training: something that is difficult, probably impossible for AI. This requires that educators are trained to integrate social interaction effectively in formal and informal education.

19. However, the Education system cannot be changed without attention to assessment and progression models. Designing progression models to underpin teaching beyond academic knowledge will require considerable human effort. However, there is a significant body of existing research that can help: helping us to specify the development of metacognitive intelligence across learners of all abilities, for example. We are perfectly capable of designing the progression models we need. We just need to put our minds to the task. We know how to define clear goals and sub-goals, to identify successful movement towards each sub-goal and goal, to provide sensitive feedback to help learners to move towards their goals and sub-goals, and we know how to help learners know how well they are doing at moving towards their goals and sub-goals. These are the components that we must now integrate to develop the next generation progression models our education systems require. For more information on the intelligence-based curriculum, please see this.[5]

20. Educators must be a core part of the 4th Industrial Revolution – we need to engage educators, technology developers and researchers in working together to develop the best AI technologies for use in education. The UCL EDUYCATE project provides a methodology for achieving this type of working.[6]

The impact of the Fourth Industrial Revolution on the delivery of teaching and learning in schools and colleges

21. The thoughtful design of AI approaches to educational challenges has the potential to provide significant benefits to educators, learners, parents and managers. But it must not start with the technology, it must start with a thorough exploration of the educational problem to be tackled. The development and adoption of AI teaching assistants[7] will provide an opportunity for developing deeper teaching skills and enriching the teaching profession. This deepening of teacher expertise might be at the subject knowledge level, or it could be concerned with developing the requisite skills to support and nurture collaborative problem-solving in our students. It could also result in teachers developing the data science and learning science skills that enable them to gain greater insights from the increasingly available array of data about students’ learning. However, whilst general funding for AI in the UK has multiplied, there is very little investment in AI for Education. This is immensely short-sighted when evidence shows that educational applications of AI can be extremely effective[8].

22. Collaborative problem-solving is a key skill for the workplace, and its importance is only likely to grow as further automation takes effect. There is currently a mismatch between the substantial evidence in favour of collaborative problem solving and learning reported in the literature and the approaches widely used within schools. Collaborative problem-solving, just like other key skill developments, does not happen spontaneously. Both teachers and students require a high level of training to employ collaborative problem-solving effectively, and yet there is little evidence of concerted training effort. This might be due to lack of resources to provide such training and support. However, AI for education implementations have good potential to contribute to training students and teachers on their key skill developments, in addition to their routine cognitive abilities. Collaborative problem-solving was considered important enough for it to be added to the OECD’s PISA assessment programme in 2015. When the results were published towards the end of 2017, Andreas Schleicher, the OECD Director for Education and Skills, urged educational systems to do better in helping their students to develop these skills (OECD, 2017).

23. The PISA results illustrated that those who perform most strongly in other PISA assessments in science, reading and mathematics also tend to perform well in the collaborative problem-solving assessment. However, the results also highlighted a cause for concern across the world. They reflected a lack of high-level collaborative problem-solving skills amongst students from all countries, including those who performed the best. Even students in Singapore struggled with the more advanced demands of collaborative problem-solving with little more than 20% of students able to attain the advanced level 4 in their PISA collaborative problem-solving assessment. This suggests that there is a great deal of work that educators need to do if we are to ensure that people have knowledge and skills that they can apply effectively when working with others in the workplace.

24. There is of course another reason why collaborative problem-solving is so important: it requires the ability to justify decision-making. This is almost impossible for machine learning systems because, while they work together, they are not able to synthesise their different domain specific intelligences and so are not able to justify their decisions.

25. Implications for teacher training and professional development: The significant educational implications that AI brings to society, both when AI is viewed as a tool to enhance teaching and learning and when AI is viewed as a subject that must be addressed in the curriculum, make it clear that teacher training and teacher professional development must be reviewed and updated. If teachers are to prepare young people for the new world of work, and if teachers are to prime and excite young people to engage with careers designing and building our future AI ecosystems, then we must train the teachers and teacher trainers and prepare them for their future workplace and its students’ needs. This is a role for policy makers, in collaboration with the organisations who govern and manage the different teacher development systems and training protocols across countries. The need for young people to be equipped with a knowledge about AI is urgent, but the need for educators to be similarly equipped is critical.

26. There are many other countries who are setting examples from which we can learn. For example: In China, Beijing Municipal Commission of Education launched the "Advanced Innovation Center Construction Plan of Higher Education in Beijing” to “integrate national, domestic and international resources, to promote both research and application, to combine technological creation and talent development, to develop both national and local colleges and universities”. The new centre at Beijing Normal University has a remit to conduct research in AI through their AITutor project drive an AI transformation of Beijing public education[9].

The role of lifelong learning in re-skilling the current workforce
Place-based strategies for education and skills provision

27. It is expected that there will be some kind of change in the practice of most professions including teaching profession due to the impact of automation. However, it is very unlikely that this impact will be too radical to replace the teaching profession, rather it is more likely that it will lead teachers to practice differently than they are today. George Bernard Shaw famously described a profession as a ‘conspiracy against the laity’. From this perspective, the impact of AI might lie in exposing the professions and diluting their power. However, if we consider a professionalism as something that marks out those who have high levels of skill and knowledge (which most people would do), then we can mainly look to AI to help us be more efficient in realising our goals, and perhaps to take the pain out of the routine elements of the teaching profession.

28. As an example, consider machine reading of student exams, are they an enhancement or replacement of professional teaching activity? If we think that the essential skill of teachers is grading student exams, then they can probably be replaced by computers who are already better at clustering, classifying and identifying abnormalities than people. However, if what teachers do is to counsel and teach students, provide them with appropriate cognitive, metacognitive, and emotional support to increase their knowledge and improve their grades, and interpret grades within their complex social contexts, then the machine-read exams will only save teachers time in the preparation for their ‘real’ profession.

29. However, such an impact will only mean that teachers will have to focus on those skills that are essential for their profession and will be spending less and less time on those skills that are prone to automation. This means that some teachers will have to be reskilled on those fundamental skills of teaching profession, but it also means that all teachers should be trained to become lifelong learners to constantly adapt into automation of the routine elements of their work. As mentioned above two key aspects that relate to lifelong learning are metacognition and self-efficacy which should be part of our education systems including our teacher training. Therefore the focus should not be on whether teaching profession will survive, but what shape it will take and how we can prepare teachers for lifelong learning.

30. Medicine and healthcare professions are an example of careers with long undergraduate, long postgraduate and significant ongoing continuing professional education and training. AI not only needs to be introduced in to the curricula at all of these levels in order to prepare these students and professionals for a workplace and a society which is changing, but AI itself may also facilitate personalisation of education to meet an individual’s learning needs.

31. In sectors such as healthcare and life sciences, a crucial part of the education strategy with respect to AI and any technology, is the ability to evaluate and to appraise evaluations of technology. An evaluation framework for AI is urgently needed at both cross-subject and subject-specific levels so that individuals can be trained to know when the use of AI is useful/helpful versus when it is neutral or harmful. This evaluation framework must be part of any education strategy for AI.  

The challenges and opportunities of the Fourth Industrial Revolution for improving social justice and productivity

32. As with every piece of technology, AI is value-laden. Therefore, these technologies have politics embedded in them. Any algorithm written in an AI system is the result of somebody deciding on a set of complex coded instructions, making various decisions, preferring some options over others based on their values. Similarly, every data set that is used to train AI systems have certain values and politics of the context from which this data was collected.

33. Therefore, every AI system somehow follows a person or a societies’ logic and values. It is our responsibility to make sure that these values that are driving AI technologies that will be implemented in educational contexts are the values we consider as significant. Arguably, one of these values should be social justice and fairness. It is important to note that every machine-based action has certain consequences and side-effects for sets of user and non-users. Some users will get more benefit from AI based decision making process than others. Considering the fact that it is extremely challenging to get over such bias embedded in AI systems (both technically and culturally), AI technologies’ limitations should be made clear to educational professionals as much as their advantages.

34. In healthcare, data is used to develop models to predict risk of disease, predict progression/outcome of disease and impact of treatments at the individual and population levels. Without representative data on grounds of gender, ethnicity, age or any protected characteristic, then there is scope for such models to exclude large sections of society and perhaps even neglect the people most in need of healthcare. Moreover, the algorithms which are used may themselves discriminate against particular individuals. Therefore, both data and algorithms must meet standards of transparency and generalisability to the population they are intended for.

June 2018