Teaching emerging technologies, AI and digital safety in schools
Pupils are using AI faster than schools can write policies for it. This page brings together what changed for UK schools in September 2026, what the research says about how young people use AI, what works when you teach it, and the Teach io resources you can use today.
- Jan 2026DfE AI safety standards updated for pupil wellbeing.
- Sep 2026RSHE 2025 and KCSIE 2026 in force, both covering AI.
- 2027Revised national curriculum due to be published.
- Sep 2028New curriculum taught, with AI literacy and GCSE Computing.
AI is now part of safeguarding and the curriculum
Until recently, AI sat at the edge of online safety. From September 2026 it is written into statutory RSHE and safeguarding guidance, and AI literacy is on its way into Computing.
RSHE 2025 guidance
Secondary RSHE now covers AI chatbots, deepfakes (how to spot them and the harm they cause), online scams, fake and AI-run accounts, online misogyny and incel culture, and the fact that using AI to create sexual images of children is a crime.
DfE, July 2025KCSIE 2026
Contact risk now includes harmful interactions with generative AI that simulates people, such as chatbots and AI companions. Every member of staff must understand their role in filtering and monitoring, and schools are pointed to the DfE generative AI product safety expectations.
DfE, July 2026Generative AI product safety standards
The DfE standards for AI tools used in education now include protecting pupils’ mental health and cognitive development: tools should spot signs of distress, signpost to human help and alert the DSL.
DfE, 19 January 2026Standard covers generative AI
The updated filtering and monitoring standard asks schools whether their systems can cope with real-time, personalised and AI-generated content, alongside revised UK Safer Internet Centre guidance.
DfE and UKSIC, June 2026Curriculum and Assessment Review
The government accepted a rebalanced Computing curriculum with AI and digital literacy for all, a broader GCSE Computing to replace GCSE Computer Science, and a possible Level 3 qualification in data science and AI. The revised curriculum is due in 2027 for teaching from September 2028.
DfE response, November 2025Education for a Connected World
The UKCIS framework sets out what pupils should know at 4–7, 7–11, 11–14 and 14–18 across eight strands, from online relationships to privacy and copyright. It is the best tool for auditing coverage across Computing, RSHE and tutor time.
UKCIS, 2020 editionPupils use AI daily, trust it easily and understand it poorly
The latest UK data shows a gap between how much pupils use AI and how well they understand it. That gap is what teaching needs to close.
of 8 to 17-year-olds use AI, rising to 66% of 16 to 17-year-olds. Many could not tell different AI tools apart and few questioned whether answers were accurate.
Ofcom, May 2026teenagers who said they were confident spotting AI-generated content could not identify the AI image they were shown.
Ofcom, May 2026of 8 to 17-year-olds say they have used AI as someone to talk to or as a friend.
Ofcom, May 2026of 9 to 17-year-olds have used an AI chatbot. 35% of users said it feels like talking to a friend, rising to 50% of vulnerable children.
Internet Matters, July 2025of children who use chatbots say they talk to them because they have no one else to talk to. Children said they want schools to teach safe use of AI.
Internet Matters, July 2025less planning time for primary teachers using Oak’s Aila (about 49 minutes a week) without lower lesson quality, but use fell over the 10-week trial.
EEF and NFER, October 2026The EEF trials concern teachers using AI to plan, not pupils learning about AI, but they carry the same lesson: tools alone do not change practice, and the EEF is now funding research into how AI affects pupils’ learning, including cognitive off-loading.
Eight research-informed principles for teaching it
Drawn from the CEOP and PSHE Association principles of effective prevention education, the UKCIS Connected World framework and the Raspberry Pi Foundation’s research on teaching about AI.
- Positive, not frightening
Effective prevention education avoids scare tactics and confrontational approaches. Factual teaching, practical skills and clear routes to help work better.
- Short, often and spiral
Interventions need adequate length and intensity. Revisiting ideas each half term, building from Year 7 to Year 9, beats one assembly or a September-only unit.
- Teach how it works, not only the risks
The SEAME framework groups AI learning into social and ethical issues, applications, models and engines. Good provision touches all four, so pupils understand why AI can be biased or wrong.
- Watch the language of ‘thinking’ machines
Research behind Experience AI warns that describing AI as if it were a person shapes pupils’ mental models. Say the model predicts or generates, not that it knows or wants.
- Check what pupils really believe
Confidence is not competence: Ofcom found confident teens still missed AI images. Assess understanding and the misconceptions behind wrong answers, not just recall.
- Whole school, with parents
Effective programmes are whole-school, led by trained staff, backed by leaders and involve parents and pupil voice. Computing, RSHE, tutor time and the DSL should plan together.
- Relevant and inclusive
Learning should reflect the platforms pupils actually use and the needs of vulnerable pupils, who are the heaviest and most trusting chatbot users.
- Review every year
Use new Ofcom and Internet Matters data, filtering and monitoring logs and pupil surveys each September to refresh examples while keeping the core ideas stable.
The SEAME framework: four ways in to AI
SEAME is a simple way to check whether AI teaching covers more than the headlines. The levels are not a sequence; strong provision touches all four.
Emerging technologies mapped to the curriculum and to Teach io
Each technology is linked to the Education for a Connected World strands, the RSHE 2025 content it supports and where Teach io already covers it.
| Technology | Connected World strand | RSHE 2025 link | Teach io coverage |
|---|---|---|---|
| Generative AI and chatbots | Managing online information; Privacy and security | AI chatbots; how data is used | Y7 L1, Y7 L3 lessons; Y7 learn guide |
| AI companions | Online relationships; Health, wellbeing and lifestyle | AI chatbots; mental wellbeing | Y8 L2 lesson; Y8 learn guide |
| Deepfakes and synthetic media | Self-image and identity; Online reputation; Online bullying | Deepfakes; AI-generated sexual imagery is a crime | Y7 L2, Y8 L1, Y9 L1, Y9 L4 lessons |
| Algorithms and recommendation feeds | Managing online information; Health, wellbeing and lifestyle | Exaggerated and extreme language; misogyny and incel culture | Y8 L3, Y9 L3 lessons |
| AI-powered scams | Privacy and security | Online scams; fake and AI-run accounts | Y8 L4 lesson |
| Bias, fairness and AI decisions | Managing online information | Stereotypes and discrimination | Y9 L2 lesson |
| Smart devices, IoT and embedded systems | Privacy and security | How data is collected and used | Y7 L4 lesson; KS3 Computing Fundamentals (embedded systems) |
| Cloud, networks and connected services | Privacy and security; Copyright and ownership | How data is used | BTEC C3 A1 Modern Technologies learn guide and assessment |
| AI, creativity and copyright | Copyright and ownership | No direct statement | Y9 L5 lesson |
| Rights and the law online | Privacy and security; Online reputation | Rules and laws on technology use | Y9 L6 lesson; KS3 impacts of technology |
RSHE links use topic headings rather than numbered statements, so check the exact wording with your PSHE/RSHE lead. Lesson codes refer to the 10-minute online safety lessons.
Teach io resources for AI and emerging technologies
Start with what is ready now. Everything below is live on the site.
10-minute online safety lessons
18 ready-to-teach lessons with AI in every one: chatbots, deepfakes, AI companions, feeds, scams, bias, sextortion and the law. Mapped to KS3 Computing and RSHE 2025.
See the lessons →Years 7, 8 and 9 online safety
Six self-marking lessons a year for pupils to work through independently, matched to the 10-minute lessons.
Open the Year 7 guide →Modern Technologies
Ad hoc networks, cloud storage and computing, platforms and services, and their implications for people and organisations, with an assessment and full feedback.
Open the learn guide →Computing Fundamentals
Impacts of technology, the digital divide and embedded systems, including the security of connected devices such as smart doorbells.
See the KS3 curriculum →Misconception-led assessment
Every wrong answer is mapped to the thinking behind it, so you find out who still believes a chatbot is a friend or a deepfake is always easy to spot.
Try an assessment →Six things to check this year
- Map RSHE 2025. Check the new AI, deepfake, scam and misogyny content is taught and where, using the Connected World strands to avoid gaps and overlap.
- Update the online safety policy. Add generative AI, AI companions and AI-generated imagery, in line with KCSIE 2026.
- Review filtering and monitoring. Complete and record the annual review, including how your systems handle generative AI, and include it in induction for all staff.
- Vet AI tools before use. Check any pupil-facing AI tool against the DfE generative AI product safety standards.
- Brief staff and parents. Share what pupils are using, how to talk about it and where to get help.
- Plan for 2028. Start building AI literacy into Computing now so the move to the revised curriculum and GCSE Computing is gradual.
Teaching the sensitive topics
Lessons on AI companions, deepfakes, extreme views and AI-generated images can lead to disclosures. Plan them with your DSL and PSHE/RSHE lead and make sure pupils know who to talk to. Help for pupils: Report Remove (Childline and IWF), CEOP and Report Harmful Content.
Emerging technologies and AI in schools: FAQ
What does 'emerging technologies' cover for schools?
For school purposes it means the technologies pupils are meeting now or will meet soon: generative AI and chatbots, AI companions, deepfakes and synthetic media, recommendation algorithms, smart and connected devices, biometrics and facial recognition, cloud services and wearables. Teaching them well means covering how they work, how they are used, and how to stay safe and act responsibly with them.
Is teaching about AI now statutory?
Parts of it are. The revised RSHE statutory guidance (DfE, July 2025) has applied since 1 September 2026 and includes AI chatbots, deepfakes, online scams and AI-generated sexual imagery. KCSIE 2026, also in force from 1 September 2026, treats harmful interactions with generative AI as a safeguarding risk. AI literacy in the Computing curriculum is planned through the Curriculum and Assessment Review, with a revised curriculum due in 2027 and teaching from September 2028.
Where should this sit: Computing, PSHE/RSHE or tutor time?
All three. The research on effective online safety education favours a whole-school, spiral approach rather than a one-off unit. Computing is the natural home for how the technology works; RSHE covers relationships, wellbeing and the law; tutor time keeps it current through the year. The UKCIS Education for a Connected World framework is designed to coordinate this across subjects.
Should we use scare stories to get the message across?
No. The CEOP and PSHE Association principles of effective prevention education specifically advise a positive approach that avoids scare tactics. Pupils respond better to factual, research-based teaching, practical skills and knowing where to get help.
How does Teach io help?
Teach io already has 18 ten-minute online safety lessons for KS3 with AI at their core, pupil learn guides for Years 7 to 9, the BTEC Component 3 Modern Technologies learn guide and assessment, and KS3 Computing Fundamentals content on the impacts of technology and embedded systems. Teach io assessments then show what pupils really believe, because every wrong answer is mapped to a misconception.
How do we keep up when the technology changes so fast?
Review examples every September using Ofcom and Internet Matters research and your own pupil voice, keep the underlying ideas stable (data, algorithms, consent, evidence, the law) and update the hooks. Short, frequent lessons are easier to refresh than a single long unit.
Research and guidance
- Ofcom, Children's Online Experiences (May 2026)
- Internet Matters, Me, Myself and AI (July 2025)
- EEF, Aila evaluation press release (October 2026)
- DfE, Relationships, sex and health education statutory guidance (July 2025)
- DfE, Keeping children safe in education 2026
- DfE, Generative AI: product safety expectations
- Curriculum and Assessment Review final report and government response (November 2025)
- UKCIS, Education for a Connected World (2020)
- CEOP and PSHE Association, Key principles of effective prevention education
- Raspberry Pi Foundation, the SEAME framework
Figures checked October 2026. Guidance changes often, so check the latest version on GOV.UK.
Find out what pupils really believe about AI
Pair the lessons with Teach io assessments. Every wrong answer points to a misconception, so you can reteach the right thing.