How AI Could Transform Education – For The Better

AI has the power to create a modern education system of the highest quality.
This post will explain how.

---

It has been just over a year since Chat GPT and AI crash-landed into our society. With this came a wave of predictions regarding its potential impact on the future of humanity. Predictions of glossy-eyed optimism around untold efficiency. Predictions of an utterly terrifying dystopia of a soulless and robotic society. Regardless of which side you land on, one prediction is certain: AI is here to stay. 

Within education, AI has undoubtedly presented some glimmers of hope. But the present reality sees educators firefighting the negative flames that have risen from the tool’s early embers. Questions linger about its effects on student engagement when seemingly no effort is required on their part to submit assignments almost exclusively produced by AI. Worse still, there are quite legitimate concerns that AI could well make the roles of teachers redundant at some point in the future. 

However, this post seeks to outline the positive case for AI in education. This technology doesn’t have to be the demise of the work of both students and teachers alike. Instead, if leveraged properly, it could be a complementary tool that pushes education to heights not seen before. It could present solutions to the various problems that plague education, from reducing teacher workload to bringing greater degrees of educational equality.

Many of the suggestions below are not yet implementable, but each one is entirely feasible given what we already know about the power of AI. 

The arrival of AI has been daunting for many of us. Yet, with the right approach at this early stage and with educators given a serious stake in its development, AI could help to produce a modern education system that allows all students to reach their very best. 

Freeing up teachers’ time to teach

A key problem facing UK education right now is the teacher retention crisis. Almost 10% of UK teachers quit across 2022. This is in no small part due to teachers routinely working a 50-hour work week, well above the OECD average of 41 hours.^[i] Currently, less than 50% of the average teacher’s time is spent working with their students, and correctly utilising AI could help staff reallocate 20-30% of their time toward actually teaching^[ii].

Reducing teacher workloads is one of the most notable areas in which AI has the potential to bring about positive change. 

That pile of unmarked papers hanging over a teacher’s head like a grey cloud? AI could be used to quickly handle this. Granted, right now AI is limited to marking multiple choice answers, but it is not far off being able to assess written student responses too^[iii].

Not only this, but AI could be capable of rapidly analysing student responses. It could highlight strengths and weaknesses, and then provide individualised feedback.

Teachers could leverage those AI insights to differentiate their lesson plans to each child’s unique needs. AI could then create scaffolds for those weaker students whilst ensuring sufficient challenge for the strongest, ranging from worksheets to whole PowerPoint presentations.

AI could enable teachers to better implement cutting-edge research into their lessons by providing summaries of the latest reports. This would make it less of a time burden to provide deeper and more up-to-date explanations to their students. 

The technology might also hold a place in more mundane tasks like sending emails, monitoring student attendance or writing student reports. 

After effectively utilising AI, teachers could find more time to engage with their students and regain their work-life balance. This could bring about the job satisfaction that would encourage them to stay in the profession.

Enhancing assessment

Teachers know the value of assessments in improving student outcomes, regardless of how much those same students may loathe having to sit them. 

But the current time burden of such assessments can leave the most valuable information to be drawn from occasional high-stakes testing. AI could open the door to more regular testing. The time burden would be reduced through automated marking, generating immediate feedback, and then providing suggestions for students on how to improve.

Imagine a situation where data could be built up regularly over a wide period of time. Gaps in student understanding would no longer be left largely to guesswork as is currently the case. Students could come to see the benefits of assessments, and rather than fear those dreaded end-of-year tests, instead appreciate the opportunity to constantly evaluate their learning. 

Individualised learning

We have all seen how AI can take our online data profiles and then provide us with sometimes scarily accurate personalised advertisements. This same automated personalisation could take each student’s educational data and personalise their learning accordingly. 

Within a class of 30 or more, despite the very best efforts of teachers, meeting the various needs of each student is often just not possible. 

Through automatically tracking and analysing student progress, AI might be able to further differentiate lessons according to varying student needs. There could be automatically generated scaffolds for those students who are struggling to grasp the content, whilst the most able are given unique extension tasks selected as per their data profile. Whilst students are working on their individualised AI tasks, this could free up teacher time to offer smaller group instruction within lessons to further differentiate learning.

Over the longer term, AI could automate retrieval practice from each student’s areas of weakness covering all their previous studies, ensuring every gap is covered. This could progress to almost unique curricula for each student based on their individual needs, allowing them to truly master a topic before moving on and thus stopping them from falling behind. 

There is also no doubt that students learn best when they can relate to the content being taught, and AI could generate content tailored to each student’s interest to ensure engagement. A class could all be learning about the same theme but through unique examples designed to appeal to each particular student. 

A unique digital learner ID^[i] for each student could contain all their educational data and allow for an immediate picture of student progress to inform the student, parents and teachers on how best to make progress. 

This digital learner ID could also be utilised to suggest tailored academic and career paths based on the individual profile of the student. All too often students make uninformed choices about their futures. Indeed, sometimes these result in happy accidents, but if we were able to ensure students were making suitable choices based on their attributes, we could allow for more fulfilling routes through education and work for each student. 

Students with additional needs could also be allowed to work more independently. For example, students with SEN or ESL could benefit from existing AI that automatically transcribes presentations delivered by teachers^[ii]. Students could also lean on AI like Grammarly for support with writing, through live feedback on grammar and language use. 

It has already been discussed how AI personalisation could make learning more engaging through personalised scaffolds for weaker students and challenges for stronger students. 

But AI can also increase engagement by allowing students to focus more on harnessing their creativity and problem-solving skills. Consider the arrival of the calculator in the 1970s. Many educators would have seen it as the end of students using their mental faculties to tackle problems. Instead, it freed students from the burden of routine calculation and granted access to tackling those elements of mathematics which require problem-solving and creativity skills. These are the exact skills which are in high demand right now and which AI could help students to develop in essay-based subjects. 

In this same vein, when researching an essay, a lot of time is spent often fruitlessly scouring long bodies of text for relevant information. In the same time that it would take to read 3 sources in depth, students could use AI to summarise 20 informational sources to give them a far superior breadth of knowledge. From here students could look instead to the task of synthesising a wider assortment of ideas. This would allow them to focus on crafting deeper arguments, and to develop those higher-order critical thinking and creativity skills that are sought after by employers. 

There is also the potential for AI to combine with virtual reality in the future. AI could make virtual reality worlds completely interactive and responsive to student input, and the potential here for learning to be engaging is almost limitless. Imagine students interacting first-hand with scientific experiments too complex (or dangerous) to be undertaken in the classroom. Or students being dropped right into the heart of the very historical events they are learning about. There is also the possibility that this situation could lend itself to the gamification of learning which is widely known to increase engagement. AI and VR together could bring learning to life in a way that only a select few of the best educators can do right now. 

There is a great deal of scope for AI to increase access to high-quality education for students from all backgrounds. 

The quality of AI education models depends heavily on the quality of their inputs. Therefore, we must ensure that the most talented teachers from across the top schools are pooling resources to create the very best AI tools. The outputs could be of the highest quality and be accessible to students who do not otherwise have access to the best schools in the country for purely geographical or financial reasons. AI in education spells the catalyst for true equality of opportunity. The outcome could be a society of highly educated people, not limited by their background. 

Now consider the implications of this on a global scale. Students in some of the poorest countries could have their learning enhanced simply by having access to a computer linked to AI programmes designed by leading educators from across the world. 

The wealthiest families know the value of specialised tutoring for their children and often spare no expense in this area. With virtual, AI-powered tutors, all children could access this very same resource at a fraction of the cost. Here we have a tool which could complete cycles by reacting instantaneously to the student’s academic data, offering suggestions for areas to study, providing scaffolded resources, assessing, and then providing feedback.

At current the progress of a student is too often hampered by SEN, language barriers, location, and economic disadvantage. With AI the effect of these factors on outcomes may be greatly diminished. 

A holistic approach to supporting students

AI chatbots are widely used by companies today to help consumers access support 24/7 without the need for human interaction. Despite their sometimes frustrating limitations, the future potential for these tools to develop into a more streamlined customer service experience is undeniable. 

This same approach is possible within education and is already being undertaken by some schools like the ‘Ed Chatbot’ in California^[i]. These AI chatbots can offer wraparound educational support around things like homework reminders, giving students and parents information on their data, live information on school transport and study techniques. If a student needs information all they need to do is ask the chatbot and it will point them in the right direction. If more specific information is needed, the bot will provide the contact details for a person in the school who can help further. And this help could be available 24/7. 

It is possible each school could have their own AI-powered chatbot to act as a one-stop shop for all queries that a student may have. And any teacher will know that students can have a LOT of queries.

‘Do I have English homework tomorrow?’, ‘What did I get in my last chemistry assignment?’, ‘When is the History trip?’, ‘How many achievement points do I have?’. All these questions could be answered at the click of a button. 

It’s fair to say that the quality of a teacher is often only as good as the training and development which they have received. Delivery of training and professional development is often left to staff in schools who have a myriad of other responsibilities to tend to at the same time. As such this vital coaching can frequently be below the standard which would ensure all teachers are reaching their potential. 

In future teacher training programmes, AI could be used to simulate students and their possible responses to teacher instruction. Teachers would have to respond to their confusion or follow-up questions, and from here the AI could give feedback on the quality of the teacher’s explanations.This practice would be conducted outside of the classroom in a low-stakes environment that doesn’t affect student learning. 

This type of AI could supplement training delivered by colleagues by offering far more opportunities for performance feedback than can be catered for by other staff who have busy timetables. 

AI could be used to suggest bespoke improvement plans for staff and help monitor their development in line with teacher standards. Again, AI outputs are only as strong as the inputs. So, any tool would require a well-thought-out design from the very best educators, selecting suitable materials which the teacher can be directed towards by the AI tool. 

There is also a wealth of cutting-edge educational research that is being produced daily that could improve teacher performance if it were truly accessible. Unfortunately, teachers do not have the time to trawl through the many research papers and much less time to then effectively implement this into their practice. AI could be used to summarise research and suggest ways in which this can be weaved into lesson delivery. 

The result of this could be better-trained teachers and subsequently the provision of higher quality education for every student. 

School inspections, and Ofsted specifically within the UK, are perhaps rightly under almost constant fire from teachers. Flying visits every few years consider only the surface level detail of a school and then brand them with a single word judgement. This assigned status can then dictate the fortunes of the school for years to come and act as a constant source of angst for almost all staff within education. 

If schools were using AI to constantly collect and analyse student data, inspectors could garner a finer-tuned picture of school performance. Schools could be compared on various metrics in real-time with one another. Using this, AI could pair up schools that perform worse in certain areas with similar profile schools that perform better in those areas. Cooperation between schools in this way could be facilitated for the betterment of all. 

Here inspections would be more of an ongoing process, that acts in real time to offer support as opposed to producing one-time judgements and leaving the school to process this over a few years. AI could be used over time to suggest improvements, taking more of the advisory process out of the hands of individual inspectors and into impartial, data-based recommendations. 

In education right now, the AI conversation is heavily orientated around the misuse of the tool by students. The most obvious means is through plagiarism, with students submitting assignments almost wholly written by AI. These are not only riddled with factual inaccuracies but can be seen as a block to student engagement with education. As such teachers are preoccupied with how to fend off this problem and detect those essays that are written almost exclusively by AI. 

There urgently needs to be effective and standardised training brought into schools to train teachers on how to limit the negative aspects of AI. Teachers also need to know how to instruct students on how to properly utilise AI and to understand the risks of misuse. 

But AI is not going anywhere and will be central to education going forward. So, as well as learning how to mitigate risks, students and staff need to be trained on how to leverage the innumerable benefitsoutlined above. Staff should know how AI can be used not only in general but specifically within their subject domain. Here, there is great scope for AI consultancies and Government agencies to support schools in making the most of this game-changing tool. 

AI in its early stages has undoubtedly caused headaches for educators as well as offering a taste of the potential benefits. Unfortunately, at present those headaches are weighing negatively on approaches to AI in education. In contrast, this post has sought to offer an optimistic outlook for the future of AI. 

Most of the ideas outlined above are not currently practicable in education. But they are all entirely possible. If educators can find a way to bring about even some of the stated possibilities of AI, then the sector could be on the precipice of a truly transformational period. 

But for this to be fruitful those same educators must have a real stake in the development of any new educational AI tools. Teachers need to force conversations on the topic within schools. School leaders need to then communicate these conversations to Governments and force them to offer teachers the space and support to be a part of the upcoming changes. AI needs to be a priority for educators and governments alike. 

Many of us live in angst at the challenges that AI could present to wider society. If we are to avoid the same fate in education, it is educators who need to take the reins in shaping its direction. Doing so could produce education systems that truly unleash the potential of our young people in a way not seen before.

---

^[i] https://researchbriefings.files.parliament.uk/documents/CBP-7222/CBP-7222.pdf

^[ii] https://www.weforum.org/agenda/2023/05/ai-accelerate-students-holistic-development-teaching-fulfilling/

^[iii] https://bernardmarr.com/how-is-ai-used-in-education-real-world-examples-of-today-and-a-peek-into-the-future/

---

^[i] https://www.institute.global/insights/public-services/future-of-learning-delivering-tech-enabled-quality-education-for-britain

^[ii] https://bernardmarr.com/how-is-ai-used-in-education-real-world-examples-of-today-and-a-peek-into-the-future/

---

^[i] https://ca01000043.schoolwires.net/Page/19423