Use AI well. Teach well with it.
If you use any AI feature on Class2Class with students, you've agreed — by accepting our Terms & Conditions at signup — to read this guide and apply it. This is how we meet our obligations under Article 4 of the EU AI Act, and more importantly, how we make sure every student on the platform has a teacher in the room who has thought about AI with care.
What we promise. What we ask of you.
Three commitments we've made to you and your students. Three responsibilities we're asking of you in return.
What Class2Class promises about AI on the platform.
- No automated decisions about students. We do not use AI to make automated decisions about students — not about grades, placements, or behaviour. Ever.
- No AI training on your data. We do not use your work, your messages, your images, or session recordings to train AI models — neither our own nor any third party's. This is contractually required in our agreements with AI vendors.
- AI identifies itself. When a feature uses AI, we tell you. Where you interact directly with an AI system on Class2Class — for example, our customer-support chatbot — that system identifies itself as AI before the conversation starts.
What we're asking of you, as a teacher using AI features.
- Know the tool. Before you use an AI feature with a class, try it yourself. Know what it does well and what it gets wrong.
- Verify outputs. AI can invent facts, misattribute sources, and reproduce bias. Check before you share or grade.
- Protect student data. Do not paste student names, emails, images, or identifying details into third-party AI tools. Not on Class2Class, not anywhere else.
Full detail is in our Ethical Guidelines §4b — For teachers.
Learning resources — pick what helps you.
Curated resources from outside Class2Class that we trust. Five topics. You don't need to work through all of them — pick the format you learn best from and the topics where you feel least confident.
Guidance for generative AI in education and research
Free PDF, written for teachers, available in several languages. Chapter 1 covers what generative AI is; Chapter 2 covers how it works at a high level.
But what is a GPT?
27-minute YouTube video, visual explanation of how transformer models actually work. Don't be put off by the length — it's the clearest plain-language explanation available.
AI Act Explained
The EU's own plain-language page on the AI Act and why it exists.
Articles on AI in education and AI limitations
A European independent research institute, high-quality writing for non-specialists.
AI Literacy for teachers
Practical guidance on how to introduce AI limitations to students and how to model critical thinking.
Day of AI resources
Short videos aimed at educators covering bias, accuracy, and classroom uses.
Guidance on processing children's data
Short, official guidance on how children's data must be handled. The directly-relevant sections are on minimisation and third-party processing.
Guidance on AI and data protection
Written in clearer English than most regulatory guidance and applies the same GDPR principles that govern Class2Class. Useful even if you're not in the UK.
Privacy and AI in Education webinars
Hour-long sessions, US-centred but the principles travel.
Ethical guidelines on the use of AI and data in teaching and learning for educators
Free, teacher-focused, EU-specific.
AI and the Future of Skills
Policy-level view of what AI literacy means for students.
TeachAI guidance videos
Led by Code.org, ISTE, WEF, and Khan Academy. Guidance videos aimed specifically at K-12 teachers adopting AI.
What to do when…
Six real situations you're likely to face as a teacher using AI with students. These are short on purpose.
…you want to use AI to draft something for your class
Use it. Then check it. Then edit it. Before you share with students: confirm facts you're uncertain about, read it out loud to catch weird phrasing, and make sure the tone is yours. If AI helped you, tell your students — model what you're asking them to do. You do not need to apologise for using AI; you do need to show your thinking.
…a student asks if they can use AI for a project
Start with: what are they trying to learn? If the learning goal is "write a short summary of this topic," AI collapses the learning. If the goal is "understand this topic well enough to explain it," AI can help them study. Decide together what's on and off limits for the specific task, and ask them to say where AI helped if they use it. There is no universally right answer — there's a right answer for this student and this assignment.
…an AI output feels biased, unfair, or hurtful
Take it seriously. Don't share it. Document what the input was and what the output said. If the tool is a third-party AI (not a Class2Class feature), try to use the tool's own "report this response" function if it has one. If the tool is a Class2Class feature, report it to us via the in-platform report button or at [email protected] — it helps us make the platform better, and other teachers will be glad you did.
…you're tempted to paste a student's work into ChatGPT or Claude to help you grade it
Don't. Even if you remove the student's name, student work can be identifying. Third-party AI tools are not covered by our contracts with vendors, and your use of them for grading creates a data protection problem that falls on you and on the school — not on Class2Class. If you want AI help thinking about grading rubrics, use it on an anonymised example, not on a real student.
…a student confides that they're using AI to "cheat" on something you haven't assigned on Class2Class
This is a conversation, not a disciplinary moment. Thank them for telling you. Ask what they were trying to accomplish and why AI felt like the right tool. Often what looks like cheating is a student who is overwhelmed or confused. The AI literacy part of the conversation — "did you check if the AI was right? did you learn anything from it?" — is the part that stays with them.
…something happens that doesn't fit any of the above
Report it. [email protected] for general concerns, the in-platform report button for anything you saw inside a project, our Anonymous Reports Form if you'd rather not say who you are. The full list of channels is in our Ethical Guidelines §10.
Ten questions to spot what to read more about.
Click any question to see the answer. We don't see your answers and we don't track who took the self-check — use it to spot gaps where you'd benefit from one of the resources above.
1 A teacher uses a chatbot to draft project feedback for a class. Which of the following is required?
Both are required. Review is for quality and accuracy; disclosure is for trust and to model honest AI use.
2 True or false: an AI feature on Class2Class may make an automated grading decision about a student.
Class2Class does not use AI to make automated decisions about students. This is a commitment, not a default — it exists even though the technology could do it.
3 A student's email and full name are about to be pasted into a third-party AI translation tool. Is this allowed under Class2Class's rules?
Student names and emails are personal data, and third-party AI tools are not covered by Class2Class's contracts. Even removing the name and using only the email would not be allowed.
4 Which of these is more likely to hallucinate a confident-sounding fact?
A language model can produce confident-sounding false statements (hallucinations). A calculator produces deterministic answers; a search engine retrieves existing content. Language models generate.
5 A student asks you whether they can use AI to write their reflection. What's a useful first response?
The question is about learning, not about AI. Starting with what the student is trying to learn keeps the conversation grounded.
6 You're preparing a lesson on migration. The AI you use gives you an impressive-sounding statistic: "93% of refugees return home within five years." What do you do first?
Don't use it. Check the statistic against a real source — UNHCR, a peer-reviewed paper, a government statistics body. If you can't verify it, it doesn't belong in the lesson. This is the single most useful habit to build with AI.
7 Which of these counts as student personal data you shouldn't paste into a third-party AI tool?
All of the above. Names, images, and writing samples can all identify a specific student. You'd be surprised how often a writing sample alone is enough to identify someone in a small school.
8 The EU AI Act requires AI systems interacting with humans to identify themselves as AI. Which article covers this?
Article 50(1). Article 4 is AI literacy; Article 72 is post-market monitoring.
9 True or false: if Class2Class introduces a new AI feature, existing teachers need to re-read this guide before using it.
The guide covers principles that apply to all AI features. That said, when Class2Class adds a new AI feature, we highlight what's new at the top of this page. Check the "last updated" date at the top.
10 An AI output contains information about a person you recognise — a public figure — that you can't verify elsewhere. You're writing a class activity and you're tempted to use the quote. What's the safer choice?
Leave it out. Don't use the quote without a verified source. The risk to a real person's reputation is larger than the benefit of the quote to your lesson.
Where to take a question.
Different kinds of questions go to different places. Here's the short version.
About a specific AI feature on Class2Class
Your first stop is the in-product help, then [email protected].
About whether something is safe to do with student data
Err on the side of caution. Ask us at [email protected].
About AI ethics more broadly
We really do encourage you to read or watch at least one thing from Section 2. Not because we want to cover ourselves — because you will be better equipped to handle something surprising in a classroom.
About something we've done — or should have done differently
Tell us. [email protected], or the anonymous reports form if you prefer.
Saw an AI output that surprised you?
Whether it was on Class2Class or somewhere else — biased, unsafe, hallucinated, or just weird — we want to know. AI feedback from teachers is how we make the platform safer for the next class.
- AI feature question[email protected]
- Student data questionGiancarlo Mena (DPO) — [email protected]
- In-platform issueThe Report button inside any project
- AnonymousDedicated form — no sign-in, no email capture
- Concern about the CEOIndependent Reviewer — [email protected]
Got feedback on this guide?
Feedback on this guide — what's unclear, what's missing, what could be better — is welcome at [email protected]. The guide is reviewed every six months and when new AI features are added to the platform.
Email feedback