1.1 Potentials & Risks
If you remember only one sentence form this page: AI has made its way into schools like a medicine that comes without a patient information leaflet. As a teacher, you don’t know when to use AI, you lack dosage instructions, and there is no information on the side effects. This chapter serves as the patient information leaflet for the Playbook.
Should AI come with a leaflet like a medicine?
You know that slip of paper tucked inside every medicine box? The one you likely toss aside without a glance? It may seem dull, but this little paper is a quiet triumph of the last century. That leaflet spells out four essentials: what the medicine treats, how much to take, the surprises it might bring, and who should steer clear.
Behind every line on that leaflet stands a whole machinery: clinical trials, regulators, and a responsibility to reveal mistakes. No one believes the leaflet itself heals you. Its real power is making the medicine safe for anyone. That is the magic we are missing when we let AI teach without guidance.
The AI tools now quietly settling into your school arrivedinherit a problem that is older than AI: educational technology has rarely come with no instructions. In the UK, a review found that only 7 out of 100 education technology companies had put their products through a controlledrandomised trial, and only 12 out of 100 had sought outside certification. TheSchools classroomrarely tellsask thefor samemore. story.In Onlya survey across 17 American states, only 11 out of 100 school leadersteachers and teachers across 17 American statesadministrators had ever requested peer-reviewed proofevidence before adopting a new tool.
The result, in many cases, is wasted money. In the US, on average two-thirds of schooleducation software licenseslicences went unused, and 98%98 out of 100 were notnever used intensively. The EdTech Genome ProjectProject, also in the US, looked at about 7,000 teaching tools worth 13 billion dollars and found that 85%85 out of 100 were either mismatcheda poor fit or misused.
This chapter may look like the dry, necessary section, but it reveals the hidden stories behind the rules. It tells what really happened, who was affected, and shines a light on the side effects, even those that quietly reached children who never signed up.
Potentials of AI
Let's make this clear right from the start. Most studies on AI's potential focus on university students. School pupils have so far featured relatively rarely in these studies. MostAnd most studies relate to short-term interventions lasting from a few hours to a few weeks. Consequently, the so-called novelty effect may have positively influenced the results. A large review of 68 experiments found exactly this pattern. The shorter the study and the smaller the number of particpants, the larger was the reported benefit of AI.
(1) Let AI explain a new topic and keep the practice for your lesson
Imagine this. Tomorrow, when a pupil misses a topic or needs that tricky concept explained yet again, let an AI tutor handle the first explanation. Reserve the practice for your lesson, where real learning comes alive.
Why this matters. In a well-controlled trial, an AI tutor taught new material at least as well as a good lesson. In a Harvard study, 194 physics students learned two topics. For one topic, they used an AI tutor developed for this purpose. For the other topic, they learned in a well-run active-learning class.class – with discussion and problem-solving in pairs, not a lecture. Each student completed both approaches, so no one can claim that the stronger group received the better treatment. In a test conducted immediately afterward, the students taught by the AI achieved significantly higher results. The researchers recorded a median time of 49 minutes to complete the task, compared with the 60 minutes allocated for the lesson in the timetable. The authors are cautious in their interpretation of these results: the material was new to the students, and much of the task simply involved explaining it well. They explicitly rejecteddecline theto ideaextend thattheir the same would applyresult to tasks requiringthat therequire integrationpulling ofseveral multipleideas ideas.together. Furthermore,And these were undergraduate students, not Year 8 pupils.eighth-graders.
A large review pooling the results of 228 studies camepoints tothe asame similarway, conclusion, except forwith one important limitation.limit. Generative AI-based chatbotsAI produced a larger effectgains onin knowing (and understanding)understanding than older forms of AI such as intelligent tutoring systems and adaptive practice software. But this was the only one of the four outcomes the review measured where generative AI had the edge.
Go deeper. Kestin, G., Miller, K., Klales, A., Milbourne, T., & Ponti, G. (2025). AI tutoring outperforms in-class active learning: An RCT introducing a novel research-based design in an authentic educational setting. Scientific Reports, 15, Article 17458. Free access, DOI 10.1038/s41598-025-97652-6. The article describes how the tutor was designed and controlled. This is useful for anyone wishing to develop an AI tutor themselves.
(2) Making better use of lesson planning time
Imagine this. Save some of the time onyou need for lesson planning by using AI to draft your first lesson plan,plan and then revise it thoroughly.
Why this matters. No one disputes that teachers’ workloads are heavy. Therefore, one potential benefit of AI is using working hours more efficiently. A UKtrial studyin England investigated the extent to which thewhether targeted use of AI for lesson planning can help to reduce teachers’ workload. Researchers randomly assigned 259 science teachers at 68 schools to either a group using a chatbot alongside brief written instructions orto a group that continuedused lessonChatGPT planningtogether aswith before.
online guide to using it well, or to a group that was asked not to use generative AI at all. Teachers in the ChatGPT group spent about 56 minutes a week planning science lessons for Years 7 and 8 (ages 11 to 13), compared with about 81 minutes in the comparison group.
A panel of subject experts assessed the quality of the lesson plans. They did not know whether the plans had been created with or without the aid of AI. The experts found no evidence to suggest that the quality of the lesson plans differed. However, two points standmust out.be highlighted. (1) The teachers used the tool less and less as the weeks went on. The researchersevaluators understoodread this as teachers moving from a learning phase into aroutine; routine. Thisit may also mean that the time saved does not occur automatically andsaving does not sustain itself.itself without support. (2) The trial measured the time spent by teachers on lesson planning, not their pupils’ learning outcomes.
Go deeper. Roy, P., Poet, H., Staunton, R., Aston, K., & Thomas, D. (2024, 12 December). ChatGPT in lesson preparation: A Teacher Choices Trial. NFER, commissioned by the Education Endowment Foundation and the Hg Foundation. Free at educationendowmentfoundation.org.uk.uk.
(3) Providing better support for learners who do not benefit fully from standard lessons
Imagine this. Your school decides how it wishes to spend its funds. It wants to purchase an AI chatbot. How likely do you think it is that the chatbot providers will provide you with evidence of the bot’s benefits for different subjects?
Why this matters. The provider may point outto research on learners with disabilities – one of the followinglongest-studied application. Chatbots are the oldest and best-documented applicationfields of ‘AI’AI in education, and thisone that is precisely what is ofteneasily overlooked when allocatingfunds funds.are allocated. In an analysis of 29 studies involving 41 groups of learners with disabilities, researchers found moderate benefits across a wide range of settings. But read the small print. In seven out of ten of these studies, ‘AI’ meant a robot:, usually a smallhumanoid physical machine sitting on a table,one, often operatedsteered by a researcher inout anotherof room.the child’s sight. Only one in five studies used software. So this is no proof that a chatbot will help an autistic pupil in your class. Nor could the reviewers say what makes these interventions effective. Based on their research, they could not identify why the studies that worked differed from those that did not.
Go deeper. Zhang, L., Carter, R. A., Jr., Liu, Y., & Peng, P. (2026). Let's CHAT about artificial intelligence for students with disabilities: A systematic literature review and meta-analysis. Review of Educational Research, 96 (1)., 215–257. Free, DOI 10.3102/00346543241293424. The real gem is the table of included studies: it instantly reveals just how little of this field actually focuses on the tools currently being marketed to schools.
(4) Teaching learners how AI fails
Imagine this. Ask your class to create ten images of 'a doctor' and ten of 'a nurse.' Tally up who appears in each set, then spark a discussion: where did these patterns come from, and who made those choices?
Fig 1. Image created with NanoBanana 2, created August, 20, 2026, One shot prompt: a doctor, a nurse
Why this matters. The classic version of this lesson feels outdated, and that is exactly why it deserves attention. For years, the demonstration was predictable: ask for a doctor, and a man appears. In onea systematictest test,by the non-profit AlgorithmWatch, Midjourney almost always pictured doctors as white men, while Adobe Firefly made visible, if imperfect, attempts to diversify. A peer-reviewed study of 2,400 AI-generated images of surgeons found the same split between products: two of three popular generators showed almost no women, while the third came close to the real proportion of women in surgery.
Try the same prompt today, and you might see a female doctor or a male nurse appear (Fig. 1). But do not be fooled into thinking the bias has vanished. A 2026 analysis of 1,344 images from three popular generators revealed the opposite: when asked for a 'competent person,' one system showed no women.
What changed is not the bias but the layer it lives in -– and thisthat is the important lesson!lesson. Human decisions stand behind any of these pictures: what the system was trained on, and what its makers afterward decided it should show. The second is adjustable, invisible, and made by a company.
- Your words are rewritten before the model receives them. From OpenAI’s own developer documentation: “we use GPT-4 to optimize all of your prompts before they’re passed to DALL-E” — and “this feature isn’t able to be disabled at the moment.” What your pupil typed is not what the machine was asked.
- Appearance gets filled in when the prompt leaves it open. OpenAI, describing its diversity technique: “This technique is applied at the system level when DALL-E is given a prompt describing a person that does not specify race or gender, like ‘firefighter.’” Afterwards, users were twelve times more likely to say the images showed people of diverse backgrounds. “A doctor” is exactly that kind of prompt.
- And that tuning can miss. Google, after pausing Gemini’s image generation in 2024: “Our tuning to ensure that Gemini showed a range of people failed to account for cases that should clearly not show a range.”
This is the lesson that deserves a student's attention, and it will outlast the next model update: the image is not a simple reflection of reality, and it did not happen by chance. Someone made a choice.
Fig. 2. Why the picture based on a prompt like "draw a doctor" looks like it does and why the picture itself won't tell you.
This lesson can start early. When 209 Finnish students in 12 classes, grades 6 and 7, explored this topic, the number who could explain the bias using data jumped from about 7 out of 100 to 4 out of 100. The honest reservation: the study measured what children could explain, not what they would do later.
Go deeper. Ali, R., Tang, O. Y., Connolly, I. D., Abdulrazeq, H. F., Mirza, F. N., Lim, R. K., Johnston, B. R., Groff, M. W., Williamson, T., Svokos, K., Libby, T. J., Shin, J. H., Gokaslan, Z. L., Doberstein, C. E., Zou, J., & Asaad, W. F. (2024). Demographic representation in 3 leading artificial intelligence text-to-image generators. JAMA Surgery, 159(1), 87–95. https://doi.org/10.1001/jamasurg.2023.5695
Vartiainen, H., etKahila, alJ., Tedre, M., López-Pernas, S., & Pope, N. (2025). Enhancing children's understanding of algorithmic biases in anand with text-to-image generative AI. New Media & Society, 27(9)., 5342–5368. Free, DOI 10.1177/1461444824125282014614448241252820. The study had no comparison group – read it as a promising classroom design, not as proof.
Weinmann, H., etMessingschlager, alT. V., & Appel, M. (2026). Gender bias in text-to-image generative artificial intelligence: Neglect and stereotypical presentations across three popular platforms. New Media & Society,Society. Online Firstfirst. 0(0)Free, DOI 10.1177/14614448261435197.
Ali, R., et al. (2024). https://doi.org/Demographic representation in 3 leading artificial intelligence text-to-image generators. JAMA Surgery, 159(1), 87–95. DOI 10.1177/146144482614351971001/jamasurg.2023.5695.
The vendors’ own statements: OpenAI (2022), Reducing bias and improving safety in DALL·E 2; Depue, W. (2023), What’s new with DALL·E 3?, OpenAI Cookbook; Raghavan, P. (2024), Gemini image generation got it wrong. We’ll do better, Google.
How solid are the research results on the potentialspotential of AI at this very moment?
The most important thing. If someone quotes a study to you, ask two questions before agreeing: "Who were the learners, and how long did the study last?"
Why this matters. In 2025, a meta-analysis of 51 studies reported a significantlarge benefit of ChatGPT for pupils’students’ academic performance. The article was read hundreds of thousands of times and served as the basis for numerous confident statements in staff rooms. On 22 April 2026, the journal retracted the article. A "retraction"retraction means that a journal officially withdraws something it has published; the article remains visible, marked with the note "RETRACTED"‘RETRACTED’, so that anyone who has cited it can recognise this. Two externalresearchers researcherswho had not been involved in the study had identified inconsistenciesdiscrepancies in the synthesis of the studies.analysis. The journal's editor wrote that the problems "‘undermine the editor’s confidence in the validity of the analysis.analysis’. The authors did not respond to correspondence on this matter.
This does not mean that AI has no impact. There are reputable studies that demonstrate moderate benefits. This field of research is also not yet well defined because it is relatively new and highly dynamic. It means that the figure that was frequently cited was incorrect, and that it took a year and two external experts to realise this.
Another reservation is worth mentioningmentioning. is:In Themathematics, at least, the closer the evidence gets to school, at least at the moment, the smaller the effect. Across subjects the picture is mixed: some recent reviews find larger effects in schools than at universities – but the school studies are still few and small.
Go deeper. The retraction itself is one page, free, and worth reading precisely because it is short: Wang, J., & Fan, W. (2 26)2026). Retraction note: The effect of ChatGPT on studentsstudents' learning performance, learning perception, and higher-order thinking. Humanities and Social Sciences Communications, 13, 528. DOI 10.1057/s41599-026-07310-z.
The retracted original was HSSC 12, 621 (2025). Still standing: Wu, X., Zhu, P., Zhang, J., Yin, M., & Wang, Y. (2026). *HSSC,ChatGPT's 13*impact on student learning outcomes: A meta-analysis of 35 experimental studies. Humanities and Social Sciences Communications, 13, 684. Free, DOI 10.1057/s41599-026-07019-z.
The school-only figure:figure for mathematics: Yi, L., Liu, D., Jiang, T., & Xian, Y. (2025). *The effectiveness of AI on K-12 students' mathematics learning: A systematic review and meta-analysis. International Journal of Science and Mathematics EducatioEducation, , 23*23(4), 1105–1126. DOI 10.1007/s10763-024-10499-7.
A broader, more recent pooled analysis: Han, X., Peng, H., & Liu, M. (2025). The impact of GenAI on learning outcomes: A systematic review and meta-analysis of experimental studies. Educational Research Review, 48, 100714. DOI 10.1016/j.edurev.2025.100714.
Risks of AI
(1) Doing something is not the same as learning
Imagine this. Before you set a homework task a chatbot can complete, think about the purpose of the task. Might it be a good idea to change the task?
Why this matters. Do you learn faster with a chatbot? Around a thousand pupils in Years 9 to 11 practiced maths in three groups: one using a standard chatbot, one using a version programmed to withhold the answer and ask a follow-up question, and one without a chatbot. Whilst using the chatbot, the group with the standard chatbot completed 48 percent more practice exercises than the group without one. Afterwards, everyone sat the same exam without a chatbot. The result, which seemed surprising at first, was that the group using the standard chatbot performed 17 percent worse than the pupils who had not used a chatbot. The group that had worked with the chatbot which withheld the answer showed no drop in performance in the test compared to the group without a chatbot, but practiced more. The pupils using the standard chatbot were neither lazy nor did they cheat. They worked harder and learnt less, because what the tool eliminated was not the effort, but the cognitive challenge that constitutes learning.
Fig. 3: More practice could result in less learning.
Go deeper. Bastani, H., Bastani, O., S ngu, A., Ge, H., Kabakcı, Ö., & Mariman, R. (2025). Generative AI without guardrails can harm learning: Evidence from high school mathematics. *PNAS, 122*, e2422633122. DOI 10.1073/pnas.24 2633122.
(2) AI Detectors punish the wrong pupils
Imagine this. You receive a student’s essay to grade. Curious about its authenticity, you feed it into an AI detector. The result flashes: the essay was almost certainly written by AI.
Why thisit matters. You use the AI detector to ensure fairness, yet for a particular group, the detectors have exactly the opposite effect. A detector does not recognise machine-generated text. It measures how predictable the wording is. And that is precisely where the problem lies, for example, when a pupil is writing in a foreign language, is unpractised in writing, or has a limited everyday vocabulary. Their writing is predictable. When seven commercialwidely used detectors analysed 91 essays written for the TOEFL English test by non-native speakers sitting– ancollected Englishfrom exam,a theseChinese essaysonline wereforum incorrectly– they wrongly flagged them as machine machine-generated in around six out of ten cases. By contrast, essays writtenEssays by 10-year-oldAmerican Americanseighth-graders writing in their native language, English,English were almost all classified correctly. A tool that triggers false positives for exactly one group is not neutral. Furthermore, this group is the least able to defend itself.
Go deeper. Liang, W., Yuksekgonul, M., Mao, Y., Wu, E., & Zou, J. (2023). GPT detectors are biased against non-native English writers. Patterns, 4(7), 100779. Free, DOI 10.1016/j.patter. 023.2023.100779.
(3) Where on earth will the pupils' data be?
Imagine this. You use an AI tool your school offers. Try to find out where the pupils' data o.go.
Why it matters. Can your school ensure that the data is not passed on to a third party? A third party is a company that is neither your school nor the app’s developer. Typically, these are advertising or analytics companies that receive data whilst the pupil is working. No one at the school chose this company. When researchers carried out a technical analysis of 163 learning products recommended by governments during the pandemic-related school closures, 145 of them processed children’s data nin a way that jeopardised or infringed their rights by transmitting data to 196 third-party companycompanies or granting them access, predomina tlypredominantly from the advertising industry. Of the 42 governments that developed th irtheir own products rather than purchasing them, 39 had developed products with the s mesame problem. The study cove scovers products from 2021, and some may have changed since then. However, we cannot rule out that the same applies to AI applications.
Go deeper. HumaHuman Rights Watch (2 22)2022). "How dare they peep into my private life?: Children's rights violations by governments that endorsed online learn nglearning during the Covid-19 pandemic. Free at hrw.org.org. The country annexes let you look up what was endorsed where.
(4) AI arrives last where it is needed most
Imagine this. How much training did you lreadyalready receive on learning about AI or subject-specific teaching with AI?
Why this matters. AdultsIn whoone hadexperiment, completedadults theirwith and without a university education weresolved provideda withwork-style business task. With an AI assistant.assistant, Theabout performancethree-quarters gof pthe gap between thosethe withtwo highergroups and lower levels of formal education narrowed significantly.disappeared. When the researchers then withdrewtook the assistant,assistant away, a sizeable part of the gap partiallycame re-emerged.back. The researchersauthors concludedconclude that the AI assistantdid hadmuch carried outof the work, butand hadthat notonly helpedpart participantsof developthe anygain skills.survived once it was gone.
A survey of American school principals revealedadds the followingsecond half of the picture: the more disadvantaged a school’s pupils were,pupils, the less impactthe school had done to integrate AI had– fewer policies, less teacher training, less guidance on teachipupil g.use. That is an association, not proof of cause.
IfTaken wetogether: combine the findings from both studies, they allow us to make the following statement: Aa good tool can bridgenarrow a gap aswhile long as iit is in the handslearner’s hands. But it tends to reach the schools that need it least, and only part of aits learner.benefit However, it usually does not reach those who need it, and leaves no skillsstays behind when it is removed.
German fig resfigures show the same pattern. Among 1,590 young people aged 14 to 20, 80 percent infrom the most affluent families andbut only 55 percent infrom the least affluent saw AI as an opportunity.opportunity Andfor recognitionschool. isUsage followfollows dthe bysame action:gradient: among young people who use AI tools, 70 percent of grammarGymnasium pupils use them for school pupils used AI foror homework, compared with 58 percent at lowerother secondary schools.
What about teacher p ofessionalprofessional development? A survey of more than t nten thousand teachers in England found that 45 percent of teachers at private schools had already received formal training on AI. ThisAI, compared with 21 percent at state schools. Within the state school system, the figuresfigure rang dranged from 26 percent at schools with the wealthiestmost schoolsaffluent intakes to 18 percent at those with the poorestleast schools.affluent.
Go deeper. Cruces, G., Fernández Meijide, D., Galiani, S., Gálvez, R. H., & Lombardi, M. (2026). Does generative AI narrow education-based productivity gaps? Evidence from a randomized experiment (NBER Working Paper 34851). A working paper, not yet peer-reviewed. Campos, C., & Singleton, J. D. (2026). AI diffusion gaps: Unequal integration of AI across K-12 schools (NBER Working Paper 35347). Latham, K., & Montacute, R. (2025). Artificial advantage? AI in the classroom and the inequality gap. The Sutton Trust. Free at suttontrust.com. Vodafone Stiftung Deutschland (2024), Pioniere des Wandels; mpfs (2024), JIM-Studie 2024. Both free, in German.
(5) What improves when learning wit AI is not necessarily what you see
Imagine thisthis.. You introduce a new ChatBotchatbot into your teaching,teaching and the pupils are enthusiastic about using the new tool.it.
Why this matters. A large review foundof fo228 rstudies looked at four different things aboutAI learningmight (Knowingimprove: knowing and understanding,understanding; workingapplying withwhat AI,one engagement,knows; motivation,the confidence,affective side, such as motivation and engagement; and the ability to plan, monitor and correct theirone’s own work).learning. Three of the four improved. The last one didn'tdid improvenot, whereasas far as the othersstudies did.could tell.
- For knowing and
understandingunderstanding, generative AIhaddidlarger effectsbetter than older forms of AI such as intelligent tutoring systems and adaptive practicesoftware that is already used in schools.software. - For applying knowledge, and for
motivationtheandaffectiveengagement,side, generative AI didnot performno betterthatthan the older software many schools alreadyexisting software.own. - For self-
steeringsteering,wethedopoolednoteffect was small and could haveenoughbeenstudiechance – and there were too few studies totellaskanywhicheffect.kind of AI might help at all.
Only a handful of the results came from randomised trials, and the enthusiasm you see in the first weeks is exactly what short studies capture best.
Here’s a quick s deside note: thein this review, the greatest impactsgains arewere found in the arts and humanities,humanities.
Go thandeeper. inYeo, mathemaG. icsH., or& theLansford, naturalJ. sciences.E. (2025). Effects of artificial intelligence on educational functioning: A review and meta-analysis. Educational Psychology Review, 37(4), article 110. DOI 10.1007/s10648-025-10085-5.
Fig. 4. Where AI might help and where to be careful (Yeo, G. H., & Lansford, J. E., 2025).
Go deeper. Yeo, G. H. & Lansford, J. E. (2025). Effects of artificial intelligence on educational functioning: A review and meta-analysis. Educational Psychology Review, 37(4), article 110. DOI 10.1007/s10648-025-10085-5.



