Tech-Inclusive Education: A World-Class System for Every Child

Today, more than ever, education must be the top priority for governments. The global technological revolution is driving rapid and irreversible changes in every area of our lives: new types of knowledge, new forms of communication, new workplace skills. Harnessing these changes, making the revolution work for everyone, is the key to economic prosperity and the central challenge for progressive politics.

With all the possibilities tech can provide, we should settle for nothing less than a world-class education for every child. Instead, education is stuck in a world of excellence for the few, not the many. The differences in outcomes both between and within countries are staggering. Like many public services, education is at a crisis point where incremental change is no longer enough.

Covid-19 disrupted the learning of more than a billion children as the resilience of education systems was put to the test. But the pandemic also showed us a way forward. All manner of technology was deployed to keep children learning. Teachers adapted rapidly and under the most difficult of circumstances to new ways of working and new ways of teaching. We saw that technology in education, far from a gimmick or a threat, is a source of resilience.

We will need every ounce of this resilience in the face of what is to come for the education sector. Analysis for this report shows the need to create places for more than 270 million more school pupils a year by 2030 – the largest expansion of school education in history. The scale of the challenge demands an immediate step change in how we think about the role of technology in education.

Brilliant examples of technology used to deliver a quality education abound. They include providing the best resources for teachers to use in classroom teaching and home learning; accurate and timely feedback to students; and engaging, collaborative learning experiences adapted to learners’ ability. But everywhere these examples are limited to pockets of good practice.

To make technology work in education, the right conditions are needed right across the board: empowered school leaders, fit-for-purpose infrastructure, better teacher recruitment and training, parental participation and above all good content, effective assessments and relevant skills for learners. Getting this baseline right is hard but crucial for technology to become more than a useful bolt-on. Systems’ starting points on this journey will differ, which is why this report proposes a new way of thinking about reform: the "minimum viable education system" framework.

With such systems in place, we can give learners control of their data and put it to use in myriad new ways, driving school improvement, closing skills gaps and helping candidates find the best jobs. The best experts in the world can enter every classroom and let teachers do what they do best, giving all children more of their teachers’ attention. We can bring new energy to the teaching profession with clear career prospects, greater autonomy and better pay. We can break the boundaries of geography with a World Education Service, free at the point of delivery, and a remote-learning infrastructure that creates true school choice without compromising children’s wellbeing or parents’ ability to go to work.

My time in government was driven by the ambition of making Britain a learning society, ready to face the challenges of the technology revolution. Today, the world itself must become a learning society. Building tech-inclusive education systems is a key step towards this.

Tony Blair
Executive Chairman

The global education crisis is real, persistent and urgent. The inequity of educational outcomes is staggering: 260 million children – a third of pupils in primary school today – will reach the last grade without learning to read; the same number are out of school entirely; in OECD countries, one in five 15-year-olds lack minimum proficiency in reading, maths or science.

We need to find cost-effective ways of delivering education of a consistently high standard to hundreds of millions of students – and the scale of the challenge, already exacerbated by Covid-19, will only grow in the next decade as the world moves towards achieving UN’s Sustainable Development Goal 4 to realise inclusive and equitable quality education for all. The solution lies in developing a radically better approach to evaluating and applying new technologies in education (edtech).

The two schools of thought currently dominant in edtech are not fit for purpose: the “tech-driven” approach, in which technology replaces or dramatically reduces the role of educators, ignores the complexities of education practices; the “tech-assisted” approach, where technology is adapted to enhance existing teaching methods, has frequently proved useful but is subject to the same logistical constraints as the rest of today’s education systems.

We propose a tech-inclusive education system in which decisions about how to use technology are made alongside and in alignment to other pedagogical choices, informed by a good understanding of their purpose and context. The goal of education policy should be to create an environment where those decisions about using tech can be freely made, and then remade as needs and circumstances evolve.

Technology offers the opportunity to deliver a better learning experience to the millions of children currently without access to the fundamentals of education: literacy and numeracy. The international community should establish a World Education Service, hosted by UNESCO or UNICEF and freely available to anyone in the world, that provides remote-learning opportunities through a range of internet and low-tech channels, and operates in close partnership with local and international edtech companies.

Unlocking the benefits of world-class education for every child requires grappling with three policy goals:

• Designing and delivering high-quality education that meets the learning needs of children and supports future economic growth.

• Meeting the logistical and systems-design challenges of providing equal access to education for all children.

• Funding the education system at a public cost that is sustainable in the long term.

The education crisis persists because these three goals are in tension with each other and the current design of school systems is not equal to resolving these tensions. To solve this trilemma, equal attention must be paid to the several key dimensions of successful education – including infrastructure, teachers, and support for parents and carers – and instead of prioritising investment in one aspect of the system at the expense of others, governments should aim to reach the baseline level across each dimension first. Our “minimum viable education system” framework lays out what a minimum acceptable level of quality looks like across each dimension in any given context.

The minimum viable education system would allow tech-inclusive approaches to education to be piloted and scaled at a national level. This would open possibilities for far-reaching reforms over the next decade that make the most of the potential of technology. In line with the framework, we recommend that policymakers should:

• Introduce a lifelong digital learner ID, issued at the start of compulsory education, that gives students control over their data, improves accountability for schools and edtech providers, and builds a comprehensive picture of a child’s learning.

• Reorganise classrooms into cohorts and transform teaching career paths, providing students with access to varied learning activities and in-person support, and allowing trainee teachers to draw on the subject knowledge of more experienced colleagues as they work in teams to deliver a mix of small-group tuition and high-quality digital learning for larger groups.

• Break the link between geography and educational opportunities and expand choice for pupils and parents by putting in place infrastructure and regulation for “in-person remote” schooling. Remote learning hubs can complement remote school places by providing supervision from trained and vetted staff, peer socialisation and a suitable working environment.

Access to education is a universal human right. Expanding access to education is also a universal good. Education transforms lives and lifts up nations. As knowledge is the backbone of economic growth, giving everyone the knowledge they require and the skills they need to make sense of the world should not be up for debate. A world-class education should be the birthright of every child.

The economic case for investment in education is as strong as the moral one. UNESCO analysis shows that an individual's earnings increase by 8 per cent for every year spent at primary school and 13 per cent for each year of secondary school. At country level, when average educational attainment rises by a year, annual per capita GDP growth increases from 2 per cent to 2.5 per cent. Ensuring all young people in low-income countries (LICs) are literate would lift 171 million people out of poverty.

Yet today, for far too many children, a good education is out of reach. Around the world, systems struggle to deliver high-quality education at significant scale and sustainable cost. There is unprecedented and unmet demand for education, with more than 270 million new secondary-school places a year needed by 2030. Funding of education is insufficient and inefficient, with just 0.5 per cent of global education spend directed to low-income countries where the need is the highest, while high-income countries face rapidly diminishing returns. There is staggering inequity of educational outcomes both between and within countries, with more than 260 million children in the Global South leaving primary school unable to read, and one in five 15-year-olds in OECD countries lacking minimum proficiency in reading, maths or science. Education systems today do not have the capacity to change at the pace required to meet the needs and aspirations of learners and best support economic growth.

Technology holds the promise of overcoming these fundamental challenges – a promise that so far remains unfulfilled.

The global expansion of access to education is one of the great success stories of the last century. A hundred years ago, only a third of the world’s population was literate. Today, 86 per cent are. Central to that expansion was the creation of functioning schools. Forged in the Industrial Revolution, the practice of compulsory schooling became a quintessential institution of 20th-century government, delivering a standardised minimum level of education at significant scale and creating robust, predictable and enduring routes to learning.

Schools also play an important broader role in our society, as the disruption caused by Covid-19 has proved. In addition to aiding the cognitive development of children, schools have a supervisory function, providing a safe environment for children during the day and freeing up parents to go to work. They provide effective structures for socialisation, helping children to develop and internalise societal values, identities, expectations around behaviour, and attitudes to citizenship and civic life. Finally, they validate knowledge and skills to support students’ participation in further education or the labour market. It is the combined impact of these roles that makes education such a powerful force for individuals, societies and economies.

But the education crisis is real and urgent. The design of school systems today, like other offshoots of industrial-era government, is fundamentally mismatched to the challenges and rapid pace of change driven by the 21st-century technology revolution. Schools prioritise consistency over responsiveness and suffer rapidly diminishing returns to scale that make access to high-quality education a matter of luck. Admitting that this is the case is not the same as saying that schools are modelled on factories or that education hasn’t changed in a hundred years, as a popular – and self-evidently false – cliché would have it. As we consider what needs to be done to fulfil the promise of universal education in the 21st century, we must consider what aspects of education systems work well and why they have not become the norm everywhere.

The answer to this last question can be expressed in terms of a trilemma. Unlocking the benefits of world-class education for every child in the world requires grappling with three policy goals:

• Delivering a high-quality education that meets the learning needs of children in a safe environment free from harassment and discrimination.

• Meeting the logistical and systems-design challenges of providing equal access to education for all children.

• Funding the education system at a public cost that is sustainable in the long-term.

Figure 1

The education-crisis trilemma

Source: TBI

The fundamental reason for the education crisis is that these three policy goals are in tension with each other and the current design of school systems is not equal to resolving these tensions. At best, they accomplish two out of three. For example, many low- and lower-middle-income countries have been able to massively scale up their primary education systems in the past two decades at relatively low cost. Yet, as the World Bank has noted, being in school is not the same as learning and the quality of education in these systems is often inadequate. On the other hand, there are some standout examples of high-quality education all around the world, but the fundamental logistical constraints imposed by the design of today’s systems mean their cost is only sustainable at the scale of a single school, school chain or perhaps, in a few cases, of a small country.

Few debates highlight this trilemma better than the vexed issue of class sizes. Research has consistently shown that one-on-one or small-group tuition produces better learning outcomes, especially for children from disadvantaged backgrounds. Reducing class sizes has therefore come to be an emotive policy issue, seen as a simple path to improving quality: hire and retain more teachers, achieve smaller classes and better outcomes. In practice, for reductions in class sizes to have an impact, they need to be significant (limiting numbers to fewer than 20 pupils in a class, and ideally closer to 12 to 15), meaning you would need to increase teaching-staff numbers in the developed world by 30 to 50 per cent (and by far more in the Global South).

Staff salaries are already the largest budget item in education, making class-size reduction prohibitively expensive at current salary levels. In the US, for example, decreasing the average student-to-teacher ratio by one would cost $12 billion a year. Reducing that outlay would make it difficult to attract qualified teachers, cancelling out any gains in quality while still incurring large costs. This is one reason why small class sizes have also become one of the selling points of private education, where resources are less constrained and student numbers are small, illustrating with painful clarity the trilemma of quality, scale and cost. It is also one of the reasons why the OECD found no direct link between class size and student performance. It has been shown that systems with relatively large class sizes, for example in Singapore or some cities in China, can perform well.

This trade-off is not unique to education, though it is particularly acute in this sphere. In many other areas, technology has shown its ability to resolve this trilemma by reducing the marginal cost of scaling identical-quality services to near zero. This makes its use in education an attractive proposition. The past decade in particular has seen significant debate over education technology (or “edtech”). Participants in this debate tend to argue for one of two approaches to edtech.

One is what we might call “tech-driven”: treating promising technological developments, such as virtual-reality headsets, artificial intelligence-driven personalisation or abundant video content, as neutral “solutions” to problems in education. In this view, technology is seen as a way to replace educators or dramatically reduce their role, and the future of education is primarily or entirely digital. This approach has been likened to putting the cart (technology) before the horse (teaching) and is increasingly out of vogue, as it fails to take into account the complexities of the learning process and the multitude of roles, not all related to teaching, that schools play. At best, this allows for reductions in the cost of large-scale delivery with a significant negative impact on the quality of learning.

The alternative, which has gained much ground recently, might be called “tech-assisted”. The argument here is that technology is at its best when it enhances what teachers already do, supporting tried-and-tested pedagogical methods or saving time. The UK government’s Edtech Strategy is a representative example of this approach. This model minimises disruption to existing classrooms and supports the use of pedagogically sound but time-consuming methods such as peer instruction, comparative judgement or diagnostic questions.

In schools that work well with technology, teachers use online platforms to find safe, vetted, curriculum-aligned content produced by their colleagues and subject experts (within the same school or elsewhere) for their lessons. These can be video lessons that bring world-class experts or senior business leaders into the classroom, polished presentations for whole-class sessions or interactive materials such as maps tagged with relevant information that students can explore by themselves or in small groups. These are the kind of materials that teachers can otherwise spend many hours preparing by themselves every week, redoing work already done by colleagues, while under significant time pressures and with fewer resources to achieve the same high standards.

Once they find the materials they like, teachers can edit content to integrate it into existing lesson plans or implement entire sequenced curricula. Less confident or skilled teachers can benefit from digital platforms that support direct instruction, a pedagogically effective approach that incorporates both learning materials for students and scripted lessons for the teachers to follow. Edtech allows for materials and lesson scripts to be rapidly updated based on evolving curricula and feedback from teachers and learners, ensuring that neither group must ever again work with out-of-date textbooks.

Interactive platforms create new opportunities to test students’ knowledge in an engaging way, to improve recall and, most importantly, to provide accurate and immediate feedback. Students compete with each other on quizzes directly linked to the learning materials and can see how well they are performing in real time. Teachers use insights from the platforms to get an immediate picture of students’ strengths and weaknesses so that, unlike with paper tests, they know exactly why a student got particular questions wrong and are able to build a picture of how their knowledge evolves over time. Edtech platforms can create individualised sets of flashcards for materials that students have struggled with so they can revise them in class or at home, while questions on future quizzes can be automatically adjusted to re-test their knowledge.

Insights from interactive platforms can also inform a more personalised approach to teaching. Teachers bring together students with similar misconceptions – or strengths – for small-group or one-on-one tuition sessions, which can also be delivered through digital platforms for after-school catch-up learning. Digital platforms can help scale “teaching at the right level” programmes, which rely on an initial assessment as well as on ongoing tests to assign students content and activities that are most relevant to their level of knowledge. This approach is particularly important in low-income settings where gaps between children’s understanding and formal curricula can be especially large. Adaptive learning platforms, which help differentiate instruction based on students’ previous performance, are also more effective for students from disadvantaged backgrounds.

Many such tools have proven successful in formal schooling environments, with leading institutions in many countries making effective use of them. However, tech-assisted education must, by definition, fit within the constraints of existing approaches and is therefore limited in its ability to resolve the crisis trilemma. Its lack of economies of scale means that the benefits of edtech are currently limited to a small subset of schools around the world: spending on education technology is estimated at less than 5 per cent of global education expenditure. Of that total, educational institutions, as opposed to parents or students, account for less than half – and less than a tenth of their digital budgets is spent on instructional or assessment software (see previous examples), even taking into account the surge in demand driven by the Covid-19 pandemic. Education's lack of digitalisation relative to other sectors means it is missing out on the opportunities that at-scale adoption of technologies creates to resolve the education-crisis trilemma.

More importantly, the pandemic has shown the limits of both approaches. We saw the failure of technology when running up against the constraints of connectivity and device access, lack of physical space for learning and the need to develop digital skills – meaning quality learning was only available to a select few. We also saw that moving traditional ways of teaching online without adapting them to this new digital context resulted in subpar learning experiences.

Learning the lessons of Covid-19 should give us a post-digital perspective on education. We need a more adaptable approach to education systems, where decisions about their design are shaped by the social and material contexts in which learning happens, and the purposes to which it is directed. These decisions should include (though not be limited to) ideas about the most effective and efficient use of technology – that is, education systems should become “tech-inclusive”. Differences in context would mean that the specifics of such decisions would vary across and even within systems. The goal of education policy should be to create an environment that minimises the barriers to those decisions being made freely – and then remade as needs and circumstances evolve.

To give every child access to a world-class education system, we must move on from the debate between “tech-driven” or “tech-assisted” approaches and become “tech-inclusive”, with a focus on conditions within the system as a whole. This, in turn, requires paying equal attention to the different dimensions of good schools (administration, infrastructure, the experience of teachers and learners, and parental engagement) and understanding what good looks like across each of them in any given context. Only then will we be able to fully address the global education crisis.

Unprecedented and Unmet Demand

A world-class education system for every child needs the capacity to absorb significant future demand on top of current pressures and to equip children for transitions between stages of education.

The right to universal access to education is enshrined in international law, and achieving it has been a key priority for several decades. In 1970, 73 per cent of children globally completed primary school; by 2009, this figure was at 89 per cent, with especially rapid progress made in the 2000s. By 2015, there were 112 million more primary-school pupils in low- and lower-middle-income countries than in 2000. And there were 57 million more children in lower-secondary schools.^[_]

One of the United Nations’ Sustainable Development Goals, SDG4, is to “ensure inclusive and equitable quality education and promote lifelong learning opportunities for all”. This includes a commitment for every child to be completing secondary school by 2030. Meeting this goal requires great ambition: every country in the world must have enough capacity in their education system for all school-age children. Globally, this equates to more than 272 million more school places in 2030 compared to 2015.^[_]

The world is not on track to meet this target. Progression through today’s education systems is linear and cumulative: completing primary education is a prerequisite for every following stage, completion of lower secondary is necessary to progress to upper secondary, and so on. In order to meet SDG4, every eligible child would have needed to enter primary school in 2018. This did not happen. Based on current trends, we estimate that only 83 per cent of secondary-school-age children will be enrolled by 2030. Yet even this relatively pessimistic scenario would require more than 160 million new school places, with at least 120 million of them in low- and lower-middle-income countries.^[_]

The challenge of growing demand for education is not limited to low- and lower-middle-income countries. Because progression is cumulative, changes in enrolment at lower levels have a knock-on effect on what happens in later stages over the next two decades. As a result, countries with high levels of school completion – typically upper-middle- and high-income countries – are facing demand pressures of their own in higher and post-secondary education. In the UK, for example, demand for university places is projected to grow by 28 per cent by 2035. Globally, by 2040, higher-education enrolment is expected to more than double.

Insufficient and Inefficient Funding

A world-class education system for every child would need to be funded at a level close to $5,000 to $7,000 per student, with funding above that level directed towards long-term transformation projects that improve resilience and equity. Alternatively, new models that produce the same outcomes at lower cost would need to be adopted.

Education as it is currently structured does not benefit from economies of scale. This means that meeting this increased demand through existing systems would require a proportional increase in education expenditure. While education spend is growing as a proportion of GDP, the growth is not fast or equitable enough. Despite the global trend towards modest increases in education spending, funding as a proportion of GDP has actually decreased in low-income countries (from 3.7 per cent in 2000 to 3.6 per cent in 2018). Meeting the target of SDG4 in low- and lower-middle-income countries has been estimated to require more than double the education spending levels of 2012, from $149 billion to $340 billion.

UNESCO estimates that global expenditure on education amounts to $4.7 trillion a year. Approximately $3.4 trillion, or a little over two-thirds, is government spending on primary and secondary education. Here, too, there are massive disparities in distribution: 48 per cent in high-income countries, 11.8 per cent in lower-middle-income countries and only 0.7 per cent (approximately $23 billion) in low-income countries. Yet 46 per cent of the world’s school students live in lower-middle-income countries and 11 per cent in low-income countries, meaning these countries are already underfunded – and they are also the ones to bear the brunt of the coming surge in demand.

Arguably, spending is not the only, and perhaps not the principal, determinant of quality in education. Countries that spend less can outperform their higher-spending counterparts on international tests of attainment, and the same amount of spending can lead to very different results. In other words, spending is a necessary but not sufficient condition of a well-performing system, and while a well-resourced system will usually outperform one that is not receiving sufficient funding, past a certain point, returns are marginal. Getting to this point will require significant upfront investment for many countries, but this will in turn allow them to ensure that operational budgets can be used to deliver high-quality education to all.

Analysis of results from the OECD's Programme for International Student Assessment (PISA) suggests that once cumulative education spend between the ages of 6 and 15 exceeds $50,000 (that is, approximately $5,500 per year, per student), additional investment has little impact on performance. This pattern is not unique to PISA participants, where wealthier countries are overrepresented. Our analysis of the World Bank’s “learning-adjusted years of schooling” (LAYS) data, which uses harmonised test scores and enrolment rates to produce an internationally comparable metric, shows that returns from extra spending diminish drastically after $5,000 to $7,000 in education expenditure by government (adjusted for purchasing parity – constant PPP) per student, per year.

About a third of the 140 countries in the data set are spending more than this, suggesting that there is funding available for long-term transformation projects to improve resilience and equity within the system without adverse impact on overall quality. Conversely, for many low- and lower-middle-income countries, reaching this level of expenditure would require increasing government expenditure by ten times or more, suggesting a need for new models of funding and cost-effective delivery that achieve genuine economies of scale without a deterioration in quality. Such models would still require uplifts to spending, both in upfront investment and in operating costs, but should help achieve results comparable to well-funded systems at a lower per-student cost. Examples such as Vietnam, Ecuador and Ukraine show that this is possible, but it is rarely accomplished today.

Poor Quality and Lack of Equity

A world-class education system for every child should ensure that all school-age children achieve at least minimum proficiency levels in core subjects, and address quality gaps faced by disadvantaged, disabled or otherwise discriminated-against students.

The question of what constitutes quality in education is not without controversy. There will likely always be disagreement over specific measures, especially when it comes to international comparisons. The “inputs” of an education system (the number of schools, learners and staff as well as the system’s financial cost) seem easier, though by no means trivial, to measure. What “good outputs” are in education is less tangible. As a complex system, education does not exist in a vacuum and there are several dimensions at play. The quality of teaching and school management, the broader school environment (infrastructural as well as psychosocial), the content and classroom materials all contribute not only to children’s academic attainment but also their wellbeing and happiness, aspirations and ability to fully participate in society.

What matter most are the specific purposes of education, and here we can identify common ground. Children should leave school with the literacy and numeracy skills appropriate to the level of education they receive. At the very least, they should learn how to read, write and count. They should acquire the social and emotional skills that are important in determining future life outcomes. Their education should prepare them to participate in society and the economy as fully as their compatriots, which also means that a quality education system is one that ensures equity regardless of gender, socioeconomic, religious or ethnic background, or disability.

On these criteria, most education systems are failing. The exact nature of this failure differs between the Global North and the Global South. In the latter, despite (and to some extent exacerbated by) the improvements in access to primary education, a shockingly large proportion of children reach school-leaving age without minimum proficiency levels in reading, writing or mathematics. To some extent, this is the result of high drop-out rates: globally, 18 per cent of children will not reach the final grade of primary school. In low-income countries, that rate is as high as 49.5 per cent (UNESCO Institute for Statistics estimate). Even those who do complete primary school, however, often leave it without having learned even the most basic reading and mathematical skills. UNESCO estimates that out of 387 million primary-school-age children who are unable to read proficiently, two-thirds (262 million) do complete primary school. Most of them are in low-income countries, where up to 90 per cent of children leave primary school without minimum proficiency in reading or maths. In such cases, although schools exist, the education system is functionally absent.

On the face of it, upper-middle- and high-income countries, which perform well on international measures such as OECD’s well-publicised PISA rankings, don’t face quality issues. However, the comparatively high average standard masks often staggering disparities in access and outcomes within individual systems. A fifth of 15-year-olds in OECD countries don’t achieve minimum proficiency levels in reading, maths or science. Pupils from disadvantaged socioeconomic backgrounds are consistently behind peers from more advantaged backgrounds in learning. In science, for example, these pupils are on average three full years of schooling behind their advantaged peers in every OECD country.

In the United Kingdom, 15-year-olds from disadvantaged families are on average 18.1 months of learning behind their peers. Children in foster care are as many as 29 months behind other children by the time they finish secondary school. In Germany, children from the top decile of socioeconomic status have the third-highest PISA score among international peers, behind only Singapore and select Chinese cities – but the mean score for students from Germany’s bottom socioeconomic decile is more than 80 points below the OECD average, equivalent to PISA’s lowest performance level and on par with the country-level scores for Albania and Peru. The gap in reading performance between the top and bottom socioeconomic deciles in Germany is equivalent to four years of schooling. The same is true of France and Israel. In the US, the difference in standardised test scores between white and black or Hispanic students amounts to approximately two years of schooling, with the gap particularly pronounced for highly segregated schools.

There are many reasons behind such disparities. To a certain extent, they are a function of the home environment and the capital – financial as well as cultural – available to parents in different population groups. But school-level characteristics such as location and pupil intake have been shown to explain a substantial proportion of the gap. Pupils in schools with a more socioeconomically disadvantaged intake perform worse in PISA tests in most countries, while disadvantaged students in advantaged schools on average perform 78 points better in their science PISA scores than those who attend disadvantaged schools.

The scale of the education crisis and the inability of today’s education systems to resolve the policy trilemma behind it suggest that piecemeal improvements will not be enough. We need to find cost-effective ways of delivering a consistently high standard of education to hundreds of millions of students. This requires a step change in our approach to the evaluation and use of technology, recognising it as an intrinsic part of the modern education process rather than something external to it. In other words, it requires creating tech-inclusive education systems.

To accelerate our journey to universal secondary education by 2030, the world must adopt a radical and pragmatic strategy that maximises the potential of technology across the short-, medium- and long-term horizons:

• In the short term (one to two years), international organisations such as UNESCO should address existing gaps in access to education by funding and building a remote World Education Service (WES), free at the point of delivery and accessible to all through the internet as well as low-tech channels like feature phones.

• In the medium term (three to seven years), national governments should build capacity for the effective use of technology through strategic investment in "minimum viable education systems", with a holistic approach to the elements of school systems that affect the rate of adoption and the impact of education technology.

• In the long term (five to ten years), national governments should pilot and scale up radical reforms that maximise the potential of technology to deliver quality education at scale and at sustainable cost. Examples of such reforms include:Introducing a lifelong digital learner ID that gives students control over their data and improves accountability for schools and edtech providers.Reorganising classes into cohorts and transforming teacher career paths to overcome the logistical constraints of traditional classrooms, providing a greater number of students at a time with access to varied learning activities and small-group tuition.Breaking the link between geography and educational opportunities, and expanding choice for pupils and parents by putting in place infrastructure and regulation for “in-person remote” schooling.

A World Education Service

To serve the needs of learners where there aren’t enough schools, where the quality of education is very poor, and in crisis situations, governments should work with multilateral organisations and private-sector partners to develop a remote education service available to all. This should focus on developing core skills (literacy and numeracy) and, where connectivity and hardware access exist, basic digital competencies. Access to content on the platform should be possible through a range of internet and low-tech channels (content preloaded onto tablets or Secure Digital cards, or delivered via hubs generating a local wireless network, or via radio and SMS). The design of this WES should be informed by the experience of the Covid-19 pandemic in low-income settings and build on the promising evidence behind the Teaching at the Right Level approach.

The WES should be hosted and supported by a main multilateral funder, such as a relevant UN agency or international development bank, but should operate as an autonomous and agile team. The team should combine technical and pedagogical expertise, and operate in close partnership with local and international edtech companies, identifying locally relevant and high-quality content suitable for remote delivery across different channels. In this way, the WES could serve as a delivery mechanism for the UN's SDG4 (Quality Education), funded primarily through international development aid. There could also be opportunities for partnerships between national governments and the WES to develop content specifically tailored to national curricula and facilitate the creation of a shared “meta-level” qualifications framework (similar to the EU’s European Qualifications Framework).

In addition to providing a backstop minimum level of education for every child, a World Education Service would provide a foundation of common experience for children from communities around the world, and opportunities to interact with learners from other countries and cultures at a young age. (Generation Global, for example, is the Tony Blair Institute′s dialogue-based global-citizenship education programme for 13- to 17-year-olds and uses remote learning to bring together students from different countries.)

Policy Implications: Next Steps

• Multilateral organisations should rapidly align existing initiatives in this space created in response to the Covid-19 pandemic, such as UNICEF’s Learning Passport (which is currently limited to a few countries), and accelerate their development, creating a global platform that can be accessed from anywhere.

• A single organisation should take control of the delivery of the WES, setting up a dedicated autonomous team to develop a simple initial assessment of children’s skills (based on the Teaching at the Right Level methodology), and engage with edtech providers to identify a focused suite of quality content suitable for digital and low-tech delivery.

Investing in Minimum Viable Education Systems

There is an urgent need for a World Education Service that serves as a backstop, mitigating the impact of failings in existing school systems. However, the broader goal should be to ensure that such failings do not occur in the first place. This requires improving the resilience of systems and their ability to fully exploit edtech’s potential to promote equitable access to a higher overall standard of education.

Edtech Is Govtech

Dreams of technological solutions to the intractable problems of education are nothing new.^[_] Over the past few decades, there has been a range of attempts made to use digital technologies to transform education. Among these, self-directed online courses, including Massive Open Online Courses (MOOCs), and hardware-focused initiatives like One Laptop Per Child (OLPC), stand out as potentially promising efforts to leverage the potential of tech to radically increase access to high-quality teaching and learning resources. These have had some impact; however, neither can be said to have solved the challenge of opening up access to quality at scale.

Despite their early promise, MOOCs have seen wide take-up but very limited completion rates and much of the take-up has been among adults who have graduated from school. OLPC was limited in its reach (in part due to cost) and not always useful to recipients because its effectiveness depended on the learners’ and teachers’ digital skills as well as appropriate infrastructure. Hardware-focused interventions often provide a similar impact to analogue interventions, including the provision of textbooks, but at a higher cost. Research has found that improving access to hardware alone does not improve test scores. There are significant variations in the impact of high or low ICT use on PISA performance between different education systems, and investing in computers for schools by itself is no guarantee of improved results.

In general, the lack of large-scale impact from edtech can be traced to a failure to consider its role within education systems rather than in relation to individual institutions or learners. Usually, edtech is seen as either a business-to-consumer (B2C) product or as a form of enterprise or B2B software, sold to individual schools or school districts. B2C edtech products – such as pre-packaged online courses or the rapidly growing category of smartphone-learning apps – depend on learners’ intrinsic motivations and require additional resources (money and time) to engage with, thus worsening existing inequalities and creating a “shadow” education system. At the B2B end of the spectrum, the market incentives push against combining scale and impact. Enterprise software achieves scale by optimising for high-volume sales processes, not educational impact. The survival of innovative startups is conditional on early traction and therefore tends to target expert users with an interest in latest practice, failing to bridge the adoption gap to the average teacher.

Breaking this cycle requires treating edtech as “govtech”, necessitating the understanding that the implementation of technology in the public sector requires a different approach than the traditional B2C and B2B models. Edtech providers need to engage closely with government at different levels, and not just the teachers and learners who use their products and services. This means recognising that there is a meaningful gap between the incentives and priorities of different parties: teachers or schools and governments on the one side, and edtech providers and investors on the other. For the former, the impact of a product on the learners (whether gains to learning or potential disruption to the classroom) is paramount, with governments also concerned about national scale. For the latter, scaling a product to national (and, eventually, global) markets is the key to long-term survival and success. Achieving impact at scale requires aligning the incentives of these two groups.

Figure 3

Illustrating the edtech priorities gap

Source: TBI

Doing so is impossible without considering what systemic changes the wider use of technology requires. In the final analysis, there are always good reasons for why an edtech product failed to scale: connectivity was not good enough; or teachers did not have sufficient motivation or time to learn the product; or learners lacked digital skills; or the training provided with the product was poor. Providers and investors are interested in improving take-up of their product, but successful adoption requires a baseline level of quality across a range of dimensions, which combine to create a positive learning environment where technology can have an impact. If only a few tech-forward schools in a system meet that level, which is often the case today, this impact will be limited in its scale. Meeting the government’s priority – system-wide adoption of best practice – requires a system-wide effort to meet this baseline. Only then will tech-inclusive systemic reform be possible, finally unlocking tech’s potential to resolve the scale-quality-cost trilemma.

A “Minimum Viable” Approach

How can we define what this might entail? In other words, what would “good” look like? In software development, there is the concept of a “minimum viable product” or MVP. With limited resources, a company prioritises in its first release a baseline level of both functionality and quality that meets the needs of its users. Launching this MVP early, rather than waiting to add a broad set of features, allows for iteration and deeper engagement with user needs while providing a good-quality initial product.

Figure 4

The “minimum viable product” approach to software design

Source: Jussi Pasanen and TBI

There are parallels here with the world of education: the challenges are significant, the resources are scarce, and the set of potential interventions is both vast and characterised by a lack of ironclad evidence. In terms of the Cynefin framework, we might characterise education as a complex domain of unknown unknowns – exactly the kind of situation that a minimum viable approach has been designed to address.

Therefore, a similar “minimum viable” approach should be adopted in carrying out education-system reform. Rather than prioritising investment in one aspect of the system at the expense of others, governments should aim to reach the baseline level across each dimension first. There is little use in giving every child a device if neither they nor their teachers have the skills to use them; the fastest connection in the world would not help to improve paper-based exams.

The “Minimum Viable Education System” (MVES) Framework

In proposing this framework for a “minimum viable education system” (MVES), we draw on research into best educational practice and input from a range of stakeholders. The framework is organised across a range of dimensions that play a role in creating a positive learning environment, enabling the successful adoption of impactful technologies. In particular, we focus on the administration of a school system at the meso level (that is, local or regional authorities or network of schools), the technical infrastructure within schools, the experience of teachers and learners, and parental or carer engagement. Because effective adoption of technology is more likely in schools where the baseline requirements of quality are met, there is significant overlap between the dimensions of our framework and good practice in school management.

Figure 5

The “minimum viable education system” (MVES) approach

Source: TBI

One obvious difficulty in proposing a single framework is that the starting point for each education system is different, and there are significant gaps between what would be considered a “minimum viable” level in low-income settings and high-income countries. For example, the levels of connectivity, class sizes and delivery capacity within the education system will differ to a very significant extent between advanced and developing economies. To address this issue, we define three tiers of the MVES:

• Tier 1 is characterised by a focus on getting the basics right. Previous research has found that poorly performing systems that improve most rapidly are characterised by higher levels of centralisation, wider use of direct instruction and scaffolding for low-skilled teachers, and a focus on fundamentals of numeracy and literacy. To make the most of technology, systems at this tier must also provide meaningful connectivity and a basic level of access to devices, which may require significant early investment. Achieving performance at this tier would represent a significant improvement for most low- and lower-middle-income countries, where only a few schools might currently meet it, while upper-middle-income countries would often perform at or near this level.

• Tier 2 is characterised by growing levels of school autonomy and a focus on continuous improvement of the quality of the teacher workforce. Other aspects include stronger accountability, including to parents, and increasing attention to non-academic elements of schooling such as careers education or extracurricular activities. Many high-income countries will operate at this tier already, although some may need to ensure performance is equal across the entire system.

• Tier 3 represents a self-improving system where schools go above and beyond their role as teaching institutions, providing support in the community and supporting the development of pupils as citizens as well as learners. Within such a system there is now sufficient capacity for far-reaching reforms and effective experimentation with technology as well as new pedagogies such as dialogue education at the national scale, without neglecting the foundations of good learning. Very few, if any, of the leading school systems will be operating at this level today, although many of the best schools in high-income countries do this to some extent.

Policy Implications

This framework can be used as both a benchmark or diagnostic tool of the current situation and as a guide to future reform. The Pareto principle would suggest that 80 per cent of schools within a system operating at the same tier might be a reasonable benchmark for meeting the tier’s minimum viable system requirements. Each tier is designed to support the implementation of the following one. As systems improve, the focus shifts from greater centralisation and minimising variation between classrooms and schools to greater levels of autonomy, peer-led creativity and experimentation, in line with earlier findings that the core drivers of performance improvement differ between “poor-to-fair” and “good-to-great” systems.

Governments in low-income countries can use the MVES approach to assess whether their strategic development plans would assist with reaching Tier 1 and whether resources are allocated appropriately between the different dimensions. International donors can also adopt the framework to evaluate whether the focus of sponsored projects (for example, school-construction efforts or careers education) would support grantee countries in reaching the appropriate minimum viable tier. For governments in high-income settings, the framework can address inequality by helping to identify both gaps in the system that require “levelling up” and areas of overinvestment. It can also inform plans for future reform.

We recognise that Tier 1, especially, represents a significant leap for many systems struggling with the trilemma of scale, quality and cost. Successful implementation would require political will and sustained focus; close engagement with local and regional authorities, school leaders, teaching unions, parents and learners; improvements to the delivery capability within systems and clear accountability frameworks; policy pilots across each dimension of the MVES and their evaluation; and partnership with private organisations as well as international ones. The upfront investment required to reach this level will be significant, and the international community should consider ways to support this via the International Finance Facility for Education, regional development banks and innovative models for private investment such as FORTE. However, we believe that successful implementation will create strong foundations for the success of societies and economies and will offer a significant prize: a world-class education system for every child.

Education Futures

The value of a minimum viable education system is twofold. First, its implementation at the right tier improves resilience and levels up standards across the system, minimising inequities in access to quality education. Second, it creates the conditions for radical, tech-inclusive reforms within the system by ensuring that all schools, rather than just a few, have the infrastructure, skills and support necessary to use technology effectively.

The final part of this report highlights examples of reforms that are predicated on the existence of a minimum viable education system. These suggestions should be approached with an iterative mindset. One of the challenges of education reform is the difficulty of building up the evidence base for transformative approaches in the context of “live” systems: the cost of failure is someone’s future. The time required for most countries to reach the appropriate MVES tier creates the opportunity to concurrently pilot the approaches proposed here at smaller scales in a range of contexts in the medium term, testing the hypotheses behind them and capturing lessons for implementation at scale in the longer term. The three proposals are:

1. Introduce a universal digital learner ID: providing students with a persistent lifelong digital record integrated with government digital ID, school information and edtech tools, giving students control over their data, and improving accountability for schools and edtech providers.

2. Reinvent teacher career paths and classroom models: removing unhelpful class-size targets and reimagining teacher career paths around in-person, cohort-based team teaching, providing a greater number of students at a time with access to varied learning activities and small-group tuition.

3. Break the link between geography and educational opportunities: expanding choice for pupils and parents by putting in place infrastructure and regulation for “in-person remote” schooling.

These proposals are interrelated: each one is feasible on its own, but if implemented alongside each other, they would support and reinforce the benefits they provide to the learners. While achieving the MVES is a precondition for national-scale rollouts, each scenario can be implemented at any tier of the framework. We hope to explore the details of each scenario’s implementation in future papers and would welcome comments and feedback from readers.

1. Universal Lifelong Digital Learner IDs

A digital ID is a core part of the infrastructure that internet-era governments should provide to their citizens to support targeted, tailored and seamless services. These IDs would typically be issued to young adults aged 16 or 18. Children should also be issued with digital IDs that serve as a consistent record of learning and that can later be converted to a “full” ID. This single identifier, issued at the start of compulsory education, should link together different data sets held by schools, edtech platforms and other providers of education-related services, putting learners in control of their data and allowing them to make better use of it.

Students must be able to grant or revoke access to parts of their record as needed. For example, they could include results of in-year tests alongside formal exam results in their university applications, or share school-attendance data as part of a job application (Estonia’s EHIS system can be used by parents to help inform decisions about their child’s education).

Even when students change schools, a digital learner ID should stay consistent and build a comprehensive picture of a child’s learning, their circumstances and their exam results. In the long term, this would become the technical infrastructure for seamless international-qualification comparability and recognition. It could also be used to seamlessly grant access to relevant public services such as free school meals or discounted transport (similarly to Scotland’s National Entitlement Card), reducing bureaucratic overheads and associated frustrations.

Education activities both require and produce significant volumes of records about learners. Today, these data are mostly used for administrative purposes, often haphazardly and inconsistently (in the UK, five different identity systems are currently used to reference records relevant to education and training). A universal learner ID, underpinned by a sufficiently strong regulatory and legal framework setting the terms of use by various parties and, crucially, rights of appeal and redress, would also transform the quality and utility of these data. Student-wellness interventions by a teacher or pastoral-care lead could be triggered, for example, if a student’s attendance dipped.

Schools could rapidly assess the impact of policy changes on students, and could support trials of new modes of teaching and assessment (Clever and Wonde are two examples of edtech companies providing single sign-on and data-aggregation services to individual schools). Governments could establish new levers of accountability and conduct near real-time analysis of gaps within the education system. The ID could also serve as an entry point for enrolment into foundational digital ID systems, where they exist. Importantly, this would be done with the full consent of the learner and in an entirely transparent way.

Policy Implications: Next Steps

• Governments should look at adapting existing digital-ID infrastructure (or projects in development) to the “learner record” use case and seek consultation on issuing digital IDs at an earlier age than is usual today.

• Edtech providers should commit to making their systems interoperable with national digital learner IDs, and collaborate with governments, educators and learners on the development of a shared schema to allow cross-border integrations.

2. Reinventing Teaching Career Paths and Classroom Models

The dominant classroom model, in which two to three dozen children take part in the same instruction-heavy activities led by a single teacher (sometimes supported by a teaching assistant), is a choice informed by logistics rather than pedagogical design, and introduces significant constraints to cost and scalability. In low-income settings, where classes of 60 or more pupils are commonplace, teachers face additional challenges in delivering high-quality instruction, pastoral support, assessment and the multitude of additional tasks teachers are expected to take on within this model.

Technology should be used to restructure traditional classrooms by introducing a blend of high-quality digital delivery to larger cohorts (for example, entire years) and in-person support. Within a cohort supported by teaching teams, learning activities could range from live broadcast or pre-recorded digital content created by dedicated teams of specialists, working independently or with an AI assistant, peer-to-peer learning, and small group or one-to-one coaching, live or over video. (At the high-performing Broadclyst Community Primary School in the UK, year groups study together in cohorts of 60, with digital instruction and in-person support from two teachers planning lessons and teaching as a team.) This would expand the range of quality content and pedagogically sound activities available to students, and provide opportunities for them to develop independent- and peer-learning skills.

With responsibility for content design and some of its delivery shared with a remote team of subject experts, in-class teaching staff, organised into small teams and working collaboratively, can step into more specialised roles, providing support one-to-one or to small groups as a subject expert or study-skills coach, supporting non-academic activities and building mentoring relationships. (In Vietnam, a successful  “high-touch, high-tech” pilot provided training and support for teachers to act as coaches alongside digital learning materials, and led to improved learner outcomes and teacher confidence).

Combined with in-service initial teacher training, this model allows trainee teachers to draw on the subject knowledge of more experienced colleagues while performing a clear set of tasks suited to entry-level skillsets. A lead teacher assumes responsibility for the team’s performance and takes on small-group tuition sessions with students to provide support for the development of learning skills. This higher level of responsibility is matched by higher levels of pay. This would help address the current challenge of limited career-progression opportunities for teachers within the classroom, comparatively low pay relative to similarly educated professionals, with little autonomy or responsibility compared to most knowledge economy workers, and attendant high rates of teacher attrition and turnover. Within low-income countries with large class sizes, this model would allow them to hire greater numbers of new teachers, shortening training timelines and ensuring they can reduce the workload for their more experienced colleagues currently dealing with classes of 60 or more children.

Policy Implications: Next Steps

• Governments should run pilots with larger cohorts accompanied by teaching teams in select schools and at different age groups (primary, lower secondary, upper secondary) to assess impact and inform initial teacher-training reforms. Inspection regimes should be adapted to allow for such experimentation, and ensure its impact on learner experiences and outcomes is consistently captured.

• For this scenario to be cost effective, it must be delivered at large scale, requiring significant investment in school infrastructure (electricity, connectivity and device availability). Governments should closely engage with existing initiatives such as Giga and accelerate their journeys towards universal internet access.

3. Break the Link Between Geography and Educational Opportunities

The widespread adoption of the internet has transformed home schooling, creating the conditions for high-quality online schools that serve the needs of children for whom in-person education is impractical or impossible. Net School and King’s InterHigh are two examples of UK online schools serving children who travel frequently or who have special educational needs that can be met with online teaching. Following the pandemic, schools are substantially more experienced at providing a remote education to their students and many plan to continue some form of online provision (more than 90 per cent of UK teachers believe edtech will play a positive role in future remote delivery).

Schools should be supported in continuing to provide online education as an option – in Canada, Ontario has taken this step. In the first instance, this would ensure that students who are absent because of health issues or difficulties travelling to school, or whose learning needs are better served by a blended approach, are included. But it could also act as a powerful catalyst for competition between schools, reduce geographic inequalities, and open up a wider range of options to students and parents.

Removing requirements to attend schools locally will open access to higher-quality schools elsewhere in a country, with institutions incentivised through funding reform to provide high-quality digital teaching. This would help equalise access to education within higher-income countries, enabling students from areas with poor-quality schools to attend higher-performing institutions elsewhere in the country. It may also incentivise local schools to improve in order to attract and retain students and funding.

The option to take up a remote school place should be complemented with facilities for play, sport and study. These “remote-learning hubs” would provide high-speed connectivity and quiet study spaces. Children enrolled in different online schools would be supervised under one roof by vetted and trained staff and socialise with each other during the school day, allowing their parents to go to work, thus removing many of the barriers that exist with home-based remote schooling. (Squirrel AI in China combines an adaptive learning-technology platform with dedicated physical locations and serves over a million learners.) These hubs could be run by community organisations, and some poorly performing schools that saw enrolment levels drop as a result of this change could be converted into such facilities.

Policy Implications: Next Steps

• Governments should consider the implications of increasing the availability of remote, school-delivered learning on existing funding formulas, and design more flexible and transferable models.

• School-inspection bodies should be adapted in response to the wider availability of remote online schooling, creating mechanisms for “digital inspection” and putting clear quality-assurance frameworks and accountability measures in place (including cybersecurity requirements).

As the ongoing global education crisis shows, school systems around the world are struggling, to different degrees, to provide high-quality education at scale and sustainable cost. This challenge is far from being solved – not only in the low-income settings where the crisis is most acute, but also in advanced economies that have largely failed to provide an equitable learning experience to their entire school-age populations despite relatively high average standards.

Technology is often presented as a potential solution to this crisis, but it is yet to have a truly large-scale impact. This is because its successful adoption is closely intertwined with other aspects of the school system. When choices about education are led by the availability of new tech solutions, they often ignore the reality on the ground, and are a poor fit for learners and teachers alike. Conversely, when technology is adopted in the service of existing pedagogical practices, it suffers from the same constraints of scale, quality and cost as the original practices themselves. What is required is an alternative approach that is informed by the specific contexts in which learning happens and the purposes to which it is directed, and where technology use is included among other aspects of the educational process: in other words, a tech-inclusive system.

Because it is shaped by context and purpose, the exact shape of a tech-inclusive system would vary between settings. However, experience tells us that a baseline standard of quality across different dimensions of the system is needed for technology to be adopted in a way that improves quality at scale and is cost-effective. In the first instance, a remote World Education Service, while in many ways inferior to a well-designed education system, would provide a “backstop” fundamental education for the millions currently without access to one. Adopting a “minimum viable” approach to the design of education systems would allow governments to ensure that reforms and funding are directed towards achieving this baseline standard and could inform a strategic approach to investment. With a minimum viable system in place, governments could leverage technology to drive transformative change within systems – better control and visibility of data for learners, accountability for schools, improved career pathways and pay for teachers, greater access and choice for learners and parents – creating world-class education systems for every child.

The authors are extremely grateful to all TBI staff and the following people for sharing their expertise:

• Alex Wong, Giga, Telecommunication Development Bureau. International Telecommunication Union (ITU)

• Beatriz Pont, OECD

• Brajesh Panth, Asian Development Bank

• Chris Husbands, Sheffield Hallam University

• Conrad Wolfram, Wolfram Research

• Jason Tan, National Institute of Education, Nanyang Technological University, Singapore

• Jonathan Bishop, The Cornerstone Academy Trust

• Julia Adamson, BCS, The Chartered Institute for IT

• Kate Dooley, Education Partnerships Group

• Kelsey Woodrick, USAID

• Les Hopper, Pearson UK

• Mary Curnock Cook CBE, independent educationalist

• Niel McLean, BCS, The Chartered Institute for IT

• Paul Morrison, Zoom

• Richard Culatta, ISTE

• Sam Freedman, Institute for Government

• Sarah Horrocks, London Connected Learning Centre

• Susan Reed, Net School

• Tom Fogden, Ada (National College for Digital Skills)

• Tristan Wilkinson, Digital Citizens

Lead Image: Getty Images