AI in education: where we are and what happens next
Throughout our 500-year history, Oxford University Press has responded to advances in the world around us to meet the needs of students, teachers, and our wider community.
2023 has seen remarkable global growth in artificial intelligence (AI) adoption, driven by the increasing availability and awareness of generative AI (GAI) tools.
When it comes to education—be that traditional classroom-based teaching or language learning—there are many considerations for what this growing interest in GAI might mean. On one hand, there are conversations about how to educate future generations on how to use AI-enabled technologies and even how to build them. On the other, there’s a pressing question around how AI-enabled technologies are positively integrated into current teaching and learning approaches. While the two scenarios often come hand-in-hand, in this report, our focus is principally on the latter. We want to look at where we are, and what will—and should—come next.
This advancing technology has huge potential in education, but GAI is not the first digital technology to disrupt education, nor even the first instance of AI making its way into the classroom. Digital resources, such as those offering adaptive and personalized learning—sometimes supported by machine learning—are already in use in many learning settings.
We firmly believe that education should always drive technology—not the other way around.
With our deep knowledge of pedagogy and understanding of the changing needs of teachers and learners, we want to develop solutions that make a real difference to learning outcomes and find the best ways to move forward on the journey together.
At the forefront of this are teachers and the need to support them with quality resources.
88% of teachers we spoke to said they would benefit from relevant insights and research being collated in the same place.
This report draws on our global network of experts, research from trusted sources, and insights from our communities of learners, school teachers, and English language teachers. We share our learnings about where we are and set out what school leaders, education bodies, and governments should consider for what happens next.
This report covers:
AI is already playing an important part in education across the globe, but how does this differ between countries?
Are teachers optimistic about AI? Do they feel prepared, or do they need more support?
Students are already using AI—how can we harness the benefits, teach the right skills, and guard against the risks of misinformation?
Many don’t have access to the internet, let alone AI—can new technology help to bridge existing divides, or will it push them further apart?
We outline five recommendations for school leaders, education bodies, and governments on how we make sure education drives technology, not the other way around.
Current approaches to AI in education differ considerably between countries. Some are treading carefully, or banning new AI tools altogether; others are rapidly embracing the technology. This reflects each country’s unique challenges and opportunities, as well as local culture, policy, and societal needs.
Educators have been a driving force behind adoption in many countries. Our own research shows that both English language and school teachers are optimistic—but cautious.
68% of school teachers in the UK and 69% of English language teachers in Europe say that they see benefits for education but are mindful of the risks.
While teachers see the potential to personalize learning and ease pressure on teacher workloads, they also see risks—including students not recognizing errors, plagiarism, and over-reliance on devices.
There are two sides to using AI in education: as a topic and as a tool. Our research shows that the topic of AI is already part of many national curricula, but the use of AI tools by both teachers and students is a much more controversial topic.
Here’s how some countries are responding to the rise of AI in education:
An AI curriculum was developed for lower secondary school pupils to be taught from September 2023. Public secondary schools are asked to teach 10–14 hours of AI as part of their ICT teaching, including ChatGPT, AI ethics, and the social impact of AI.
Schools are piloting an interdisciplinary approach, with AI learning embedded in subjects including ICT, science, maths, language, and social studies.
AI is an optional area alongside other skills subjects at secondary level. The National Curriculum Framework for School Education, released this year, suggests that teachers and students can use AI tools to enable learning. Initiatives include Responsible AI for Youth (YUVAi): a national programme for government schools to develop a new-age tech mindset and relevant skill sets.
AI is taught as part of Robotics as an after-school activity in primary education, and as part of the Technology curriculum at secondary level. Spain is also part of AI+, an initiative coordinated by the Universidade da Coruña, which aims to introduce AI into the secondary school curriculum across Europe.
The Department for Education (DfE) has requested feedback from teachers and industry experts to understand how AI should be used in the classroom. Possible outcomes include guidance for students on the best ways to use GAI, and further exploration of AI’s potential to save teachers time, and so alleviate workload pressures and the recruitment and retention crisis. Meanwhile, a global summit on AI safety is taking place in the UK in November 2023.
Until recently, education authorities and schools were divided over the use of GAI in schools. The use of GAI was subsequently banned in public schools across several states temporarily. As a result, some commentators challenged whether this was in students’ best interests.
An Australian Framework for Generative AI in Schools was released for consultation in mid-2023 to guide teachers and students on how to use GAI tools in a safe, responsible, and ethical way. This has now been approved by all education ministers. AI, including ChatGPT, will be allowed in all Australian schools from 2024.
- While countries are taking different approaches to exploring AI—and in particular GAI—there’s no denying that it’s prompting questions about the future of education across the world.
- It’s a constantly changing picture and one that will no doubt look very different this time next year, but we can be sure that differences in approach will continue, potentially resulting in different outcomes. While some may be better placed to take advantage of new opportunities, others may miss out on key skills that could worsen problems like misinformation. And some may not have access to such tools at all, further exacerbating the digital divide.
More than 70% of teachers are optimistic about the role of AI in education.
AI has been hailed as a possible solution to many challenges in education, and we’re only just beginning to see its potential to alter how education is delivered.
Currently, teachers spend a significant amount of time devoted to work that happens outside of the classroom—whether that’s evaluation and feedback or preparation. And with teachers shortages around the world on the rise, the integration of AI-enabled tools could transform teachers’ workdays, making extra time for teacher-led instruction and live interaction.
Where are we now?
We asked our global network of school and English language teachers to share reflections on the impact AI is having on them right now: 38% of school teachers say they have used AI-based tools in their teaching in the last year. For English language teachers, it was 36%.
Most teachers think digital resources—including AI-powered technology—have had ‘significant’ or ‘some’ positive impact on educational outcomes for their students (63% of school teachers, 59% of English language teachers).
The biggest impacts so far are on content preparation, assessment, and personalization.
On the value of adaptive learning and ongoing assessment, one teacher told us:
“AI-powered adaptive learning platforms can dynamically adjust the difficulty level of content based on students’ performance and provide real-time feedback. This approach ensures that students are appropriately challenged and can progress at their optimal pace, enhancing their engagement and understanding of the material.”
– School teacher, Pakistan
Building on previous experience
Our earlier research shows that use of digital resources increased rapidly during the Covid-19 pandemic, with 36% of education institutions using digital platforms by the end of the first year and 45% of learners accessing digital learning from home. As face-to-face learning returned, digital has remained an important tool in the classroom to help achieve successful outcomes.
“Online testing on demand offers students the chance to perform in a Wi-Fi connected classroom or at home. Various digital resources can be used in teaching and learning for the purpose of active learning. Language teachers must continuously learn new technology including AI apps to teach effectively and attractively.”
– English language teacher, Japan
Teachers need help preparing for AI
While teachers are cautiously optimistic about AI in education, they need more help to prepare for the impact AI-enabled tools will have in the classroom—especially in the UK. Only 23% of UK teachers felt prepared, compared to more than half of those based outside the UK. Around a third of English language teachers felt prepared for AI, with another third not feeling prepared and the final third unsure either way.
Although there have been many exciting advancements in GAI, its adoption in the classroom should always be driven by educational requirements.
Advice to teachers from our expert speakers included:
“Be clear on your learning goals and the role that technology can play in supporting them.”
– Wayne Press, Global Product Director, Education Division, OUP
“The more that educators can do to try and assess whether these tools are actually making a difference to the efficacy of teaching—and then sharing that knowledge—is really key.”
– Professor Victoria Nash, Director of the Oxford Internet Institute
Many educational and academic institutions, including the UK’s Russell Group of universities—of which the University of Oxford is a part—have already produced their own principles on ethical usage to guide their members and educators in general.
“Teachers need tools they can trust, that are simple to use, and that will help them free up more time for the irreplaceable moments of human interaction between teacher and student. We’re excited to be developing resources that make the most of the opportunities offered by AI, whilst keeping teachers at the very heart of English language teaching.”
– Sarah Ultsch, Product Strategy Director, ELT Division, OUP
- The adoption of AI tools by teachers should have a clear educational focus. More support, through training or by having dedicated AI leads, must be given to educators on using AI-enabled tools in a way that is both ethical and beneficial to learners. This will help to mitigate ad-hoc experimentation and keep a focus on learning goals. In order for this support to be effective, it must be a joint effort between governments, institutions, and educational resource providers.
- Teacher preparedness is a persistent challenge. In 2021, our Addressing the Digital Divide report showed that over half of teachers and learners globally lacked the skills needed to make digital learning a success. As a result, we made recommendations for building digital competencies. The rapid advancement of GAI makes this more urgent and adds the additional complexity of getting comfortable with a new type of technology.
During our outreach to our communities, we spoke to our communities of school and English language teachers to hear about the impact AI is having in their classrooms—on both themselves and their students.
How many students are already using GAI to do their homework?
Almost half (47%) of school teachers in the UK told us that they suspect their students are already using AI tools such as ChatGPT.
But within English language teaching, the picture was very different. Only 20% of teachers in Europe think their students are using AI tools, and only 16% of teachers in Latin America.
This follows a recent survey by RM Technology which found that two thirds of teachers in the UK believe they are regularly receiving work written by AI. Learners also claimed to find the tools beneficial, with 68% attributing better grades to AI, and nearly half (49%) warning that excluding AI would have a negative impact on their learning.
In another survey with our teaching community, 62% told us that they felt AI poses a high risk to the integrity of students’ work in homework and assessments, with a further 32% feeling it represents a medium risk.
Teachers we spoke to remain divided on whether AI can have a positive impact on educational outcomes.
Some acknowledged the benefits it could have for learners:
“AI-powered technology in education holds great promise, and as a cautiously optimistic individual, I believe it can bring about transformative changes. Personalized learning experiences enabled by AI can improve student outcomes and engagement by tailoring content to individual needs.”
– school teacher, UAE
“If it is used properly, it can only have a positive impact on student learning, creating individualized lessons for all types of learners, and not a ‘one for all’ methodology. I see AI as the ultimate teaching tool for the learner, where the teacher creates materials, questions, or activities for the learner to then use AI to ‘assist’ them in their learning path.”
– English language teacher, Italy
“Even though I have not fully explored the possibilities of AI, my first-hand experience with it is mainly positive… Students can learn how to create prompts to make the most of AI, use it only as a source of information (not as a final product), and then interpret it from their points of view.”
– English language teacher, Bulgaria
Others highlighted the potential risks of students using AI and emphasized how it shouldn’t be used as a substitute for some of the most important parts of learning:
“Overall, I see more negatives as students depend too much on technology. They are unable to think critically on their own and cheating is prevalent when technology is used in class.”
– school teacher, Zimbabwe
“I’m afraid that students are going to use [AI] to get out of putting the work in when it comes to doing their homework, and that it’ll add to my workload to check if their assignment might have been completed by using AI technology.”
– English language teacher, Hungary
“It needs to be used with caution, like with every other technology. It makes you reassess the point of knowledge acquisition when AI removes the need to be specialized in a particular topic. Perhaps students will need to be taught analytical and critical skills—more so than recalling facts and figures?”
– school teacher, Australia
Skills for the future
As the world continues to experience rapid social, economic, and technological changes, the discussion around the skills that future generations will need to succeed continues to evolve.
While 84% of our teachers agree that it is very likely learners will need AI skills to succeed in the future workplace, they rated their schools’ current curricula and teaching methods a 5 out of 10 for how well they prepare for a future with AI technology.
It’s clear that AI is going to have—and is already having—a profound impact on the world of work. The World Economic Forum anticipates that the percentage of tasks completed by automation will increase from 34% in 2023 to 43% in 2027.
This suggests the uniquely human skills that are harder for technological systems to replicate, such as critical thinking, will be vital as our world becomes even more technologically driven.
UNESCO’s latest guidance for policy-makers also emphasizes the importance of these skills. Their report warns that AI shouldn’t be used where it can impact the development of cognitive and social skills gained through real-life observations, interactions with other human beings, and practical experimentation.
One major challenge presented by the growth of AI in wider society is the rise of misinformation.
It seems likely that AI-generated misinformation will become just as prevalent online as human-generated misinformation is today—and potentially even harder to spot, according to one study. There is also significant concern around the ways in which information delivered by GAI tools could amplify existing bias and stereotypes.
Sam Altman, CEO of OpenAI, warned that the exploitation of AI technology could seriously impact key events such as election cycles.
Our own research showed that young people are increasingly going online to source information. With the rise of social media, people are now exposed to associated misinformation from a much younger age. AI threatens to exacerbate this issue, making it more difficult to separate fact from fiction.
“It’s important that education continues to evolve to provide children with much-needed critical thinking skills from a young age to teach them to examine the information they encounter and reach their own conclusions about its validity…helping children use technology in a safe and responsible way, and equipping them with the tools to be well-versed in digital literacy for the future, is crucial.”
– Fathima Dada, Managing Director of OUP Education Division, OUP Matter of Fact Report, June 2022
The misinformation challenges around AI make education even more important. A survey by EdWeek Research Center showed that 72% of educators felt it was important to teach students to use AI-driven technology effectively and understand the pitfalls.
- AI is an inevitable part of learning, life, and future work for today’s learners. Education needs to support them to use it effectively while ensuring they benefit from an education that teaches the right skills and is rooted in quality resources and information.
- There are undoubtedly great opportunities in the wider adoption of AI-powered technologies, but this should be driven by the positive impacts it can have on the learning experience, rather than the technology simply being available.
- While this will undoubtedly be challenging to navigate, it will also bring about new benefits like the creation of new job roles, and the opportunity to focus on core skills such as creative thinking.
- We must equip young people with the necessary skills to use these new tools safely and ethically, including being able to identify reliable content over misinformation.
Technology adoption is never uniform—and there’s always a risk that disadvantaged educators and learners will be left behind. In many places, access to AI is a secondary issue; reliable access to the internet must come first.
Our research into the digital divide showed that poor digital access (physical access to the internet or a digital device) was the biggest barrier to digital learning—a problem identified by 68% of teachers. A lack of digital competency among learners and teachers compounded this issue, with over half (56%) saying skills gaps were a barrier to effective digital learning.
In some cases, AI can be used to level the playing field for learners. In Singapore, some teachers are using AI tools to identify knowledge gaps and put tailored intervention in place, supporting low-progress learners and those with diverse learning needs. This points to the potential AI has for addressing one-size-fits-all approaches to education—something that we know is never possible due to individual strengths, weaknesses, preferences, and interests.
But without adequate infrastructure in place to provide equitable access for all, there is a real danger that AI could exacerbate existing inequalities. As a teacher in Ghana told us, “…there is a potential challenge, particularly for students in Africa with regard to access to computers, laptops, and phones for AI usage.”
If teachers and learners can’t access the technology in the first place, then they won’t be able to experience any of the benefits—nor keep up with the wider global changes AI brings.
“We are struggling to align the speed of transformation of the education system to the speed of the change in technological progress.”
– Stefania Giannini, assistant director-general for education at UNESCO speaking to Reuters
- The uneven rollout of AI across education risks exacerbating the digital divide.
- As recommended in our Addressing the Digital Divide report, governments should target resources to address both ends of the digital divide, from investing in affordable access to reliable internet connections and devices to creating a focus on digital competency for teachers and learners.
- Even in regions with reasonable digital access, anticipated teacher shortages mean there is a risk that cheap or free AI tools are seen as an adequate substitute for teacher-led instruction and the acquisition of skills and knowledge.
- To avoid exacerbating existing divides, there should be a clear educational focus for the implementation and use of AI-powered technology where it can be accessed.
63% of school teachers and 59% of English language teachers we surveyed believe that digital resources, including AI technology, have already had a positive impact on educational outcomes for their students.
Meanwhile, a report from the Capgemini Research Institute found that over half (56%) of teachers globally agree that curricula and assessments need to adapt to the emergence of AI. This was as high as 70% in the US and 69% in the UK, but only 36% in Japan and 24% in Singapore, reinforcing what we have seen in this report: adoption and attitudes are not consistent across the world.
“In order to unlock the potential of technologies like AI, it’s really vital that we think carefully about the decision that we take when we’re using these tools…They should be seen as part of these overarching goals we have for education and our uses should be mindful and critical of that”
– Professor Victoria Nash, Director of the Oxford Internet Institute, speaking at the Oxford Forum, April 2023
”AI technologies have the potential to provide fantastic learning solutions for everyone in the English language learning community, as long as their application remains rooted in quality and expertise. Our focus for any resource we publish is always to help more English language learners achieve better educational outcomes, empowering them to make progress on their language learning journey.”
– Santiago Ruiz de Velasco, Managing Director of OUP’s English Language Teaching Division
Different policies and approaches across the world, as well as the speed of change, mean that we will undoubtedly see a necessary period of experimentation and innovation. However, without government support to establish principles and regulation, this could lead to a dip in the quality of education provision.
To avoid this, we conclude our report with recommendations for the consideration of school leaders, education bodies, and governments:
- Support teachers, don’t substitute them – AI should help teachers to perform their role rather than seeking to replace them. When implementing AI-powered technology to aid productivity and deliver personalized learning opportunities, school leaders, education business leaders, and policymakers must continue to prioritize the role of the teacher as guide, advisor, and supporter.
- Seek the highest quality resources – More than ever, resources created and curated by education
publishers play a critical role in maintaining the highest standards and most impactful learning outcomes. Turning to free AI tools for classroom content generation risks diluting quality and disturbing the sequence that learners need to succeed. Governments should consider how regulation can help to maintain standards in teaching and learning resources during the coming period of experimentation and rapid innovation.
- Empower teachers to use digital technologies in the classroom – With workloads so high and changes so rapid, governments and school leaders should provide comprehensive support and guidance to teachers to build their confidence, knowledge, and skills when using digital technologies in the classroom. This could include dedicated educational leads in AI and quality training resources.
- Equip students with the necessary skills to complement GAI – The potential that some AI tools have to produce misleading or inaccurate information underlines the need to embed uniquely human skills—critical thinking, creative problem-solving, and digital literacy—across the curriculum, at every stage of the learning journey and in every subject. Regulation of AI tools, particularly in educational settings, is necessary to protect students from the risks of misinformation.
- Prioritize genuine topic understanding – Where free AI tools are positioned as providers of knowledge, they naturally encourage students to bypass processes like practical experimentation and independent cognitive reasoning by which they actively build their understanding of different subjects. We must adapt curricula and assessment to keep exploration, reflection, and interaction at the heart of the learning experience.
”We need to be cautious of the behaviours that hype drives… The fundamental flaw is when we use these technologies without thinking about it as part of a broader pedagogical plan and really what we’re trying to achieve from a learning outcome perspective.”
– Wayne Press, Global Product Director for OUP’s Education Division, speaking at the Oxford Forum, April 2023.
About our research
Our teacher community is made up of more than 8,000 educators across the world. This report is based on research by 1,280 responses from our teacher communities across the world who we asked about their views on AI and future skills. This was supported by a review of current research and expert commentary on AI in education, with thanks to Green Shoots.