# AI in schools and colleges: what the DfE expects, and protecting assessment

*What staff may use AI for, what must never go into it, and how to keep coursework honest.*

## Production summary

- Modules to record: 2
- Total script: 892 words, about 6 minutes of finished audio
- Voices: Emma (host) and George (practice educator)
- Level: Teachers, teaching assistants, tutors, exams officers, school leaders and governors

## Accreditation wording that must appear in the description

- **The CPD Certification Service** (planned): Application scheduled.
- **Department for Education: generative AI in education** (aligned): Written against the DfE's published guidance and product safety standards. Our own mapping, with no endorsement from the Department implied.
- **Joint Council for Qualifications: AI use in assessments** (aligned): The assessment material is written against JCQ's published guidance. Our own mapping, with no endorsement from JCQ or any awarding body implied.

> Do not upgrade any of these words in a description or a thumbnail. Aligned is not accredited, and planned is not approved.


---

## Using AI as a professional: what it is for and what it must never be given

**Runtime** about 3 minutes. **Words** 429. **Starts at** 00:00 in the full course recording.

### Learning outcomes to state on camera

- State what the DfE says staff may use generative AI for
- Explain where responsibility for the output sits
- Decide what must not be entered into an AI tool
- Explain the rules on pupils' use and supervision
- Describe what the product safety standards expect

### Script


`[CUE 1]` *A teacher's desk with four permitted uses: planning, resources, feedback and admin, and the teacher checking the result*

**EMMA**  [00:00]
George, half my department use AI to plan lessons and the other half think it's cheating. Who's right?

**GEORGE**  [00:07]
The Department for Education is on the side of the first half, with conditions. Their guidance on generative AI in education names lesson and curriculum planning, creating resources, tailored feedback and admin as things staff may use it for.

**EMMA**  [00:22]
So it's fine.

**GEORGE**  [00:24]
It's allowed. And the same guidance says the quality and content of any final document remains the responsibility of the professional who produced it. A worksheet with a factual error in it is your worksheet, whatever drafted it.


`[CUE 2]` *A list of school information stopped at a gate: names, results, special needs, family details, and safeguarding marked never*

**EMMA**  [00:39]
So I check it.

**GEORGE**  [00:40]
Facts, level, tone and bias. Before it reaches a pupil or a parent. The tool can draft. You're still the teacher.

**EMMA**  [00:49]
What can't I put in?

**GEORGE**  [00:51]
The guidance recommends personal data isn't used in generative AI tools at all. In a school that's a lot. Names, dates of birth, results, behaviour records, special educational needs, anything about a family.


`[CUE 3]` *A pupil's essay with a small ownership mark, and a permission slip from a parent*

**EMMA**  [01:04]
And safeguarding?

**GEORGE**  [01:05]
Never. Not anonymised, not summarised, not to tidy up a record of concern. That one has no exceptions.

**EMMA**  [01:12]
I was going to paste a set of essays in to help with marking.

**GEORGE**  [01:18]
Then here's the rule that surprises people. A pupil's original work belongs to the pupil. The guidance says schools must not allow it to be used to train an AI model without permission. And permission means the pupil's, or a parent's or guardian's.


`[CUE 4]` *A classroom using an AI tool under supervision with an age rating and a filter*

**EMMA**  [01:35]
I'd never have thought of an essay as intellectual property.

**GEORGE**  [01:39]
It is. And whether pasting it in trains the model depends on the tool's terms, which is why you use tools the school has approved.

**EMMA**  [01:49]
Can pupils use these tools themselves?

**GEORGE**  [01:51]
Only with safeguards, the guidance says. Close supervision, and tools with safety, filtering and monitoring features. And schools must comply with the age restrictions the tools set. Many general tools aren't intended for children.


`[CUE 5]` *Four questions on a card for choosing a tool, and a school's approved list*

**EMMA**  [02:05]
How do I know if a tool is safe for a class?

**GEORGE**  [02:10]
The Department publishes product safety standards for generative AI in education, updated in January 2026. Filtering harmful content, monitoring and reporting, security, privacy and data protection.

**EMMA**  [02:20]
That's for suppliers, surely.

**GEORGE**  [02:22]
And for whoever chooses the tools. But it gives you the questions. Does it filter? Does it tell us when a pupil hits a filter? What happens to what pupils type? Has the school approved it for this age group?

**EMMA**  [02:38]
And the one thing to do tomorrow?

**GEORGE**  [02:40]
Ask for your school's list of approved AI tools and what each may be used for. If you're using one that isn't on it, stop and ask.

### Sources for the on screen credit

- Generative artificial intelligence (AI) in education (updated 12 August 2025), Department for Education
- Generative AI: product safety standards, Department for Education
- Using AI in education settings: support materials, Department for Education
- Keeping children safe in education, Department for Education

---

## Assessment and authentic work: the rules, and what to do when you suspect misuse

**Runtime** about 3 minutes. **Words** 463. **Starts at** 02:51 in the full course recording.

### Learning outcomes to state on camera

- State what counts as AI misuse in assessed work
- Explain what a student must acknowledge and keep
- Authenticate a pupil's work
- Explain why an AI detection tool is not proof
- Act correctly when misuse is suspected

### Script


`[CUE 1]` *A coursework folder with the rule on its cover: the work must be the student's own*

**EMMA**  [02:51]
George, I've got a piece of coursework in front of me that doesn't sound like the boy who wrote it. Where do I start?

**GEORGE**  [03:01]
With the rule. The Joint Council for Qualifications publishes guidance called AI Use in Assessments, most recently on 30 April 2025. The starting point is that students must submit work for assessment which is their own.

**EMMA**  [03:15]
So any use of AI is cheating.

**GEORGE**  [03:18]
No. Misuse is where a student uses an AI tool and submits work that isn't their own, or doesn't acknowledge it. Copying or paraphrasing AI generated content so the work's no longer theirs. Or using it as a source and not saying so.


`[CUE 2]` *A student's acknowledgement: the tool's name, the date, a screenshot of the questions and answers, and how it was used*

**EMMA**  [03:35]
And that's malpractice.

**GEORGE**  [03:36]
It is. And the sanctions include loss of marks, disqualification, and being barred from entering for a period. Students need to be told that before they start.

**EMMA**  [03:47]
What if a student used it honestly?

**GEORGE**  [03:50]
Then they show their working. The name of the AI source and the date the content was generated. A copy of the questions they asked and what came back, in a form that can't be edited, like a screenshot. And a brief explanation of how they used it.


`[CUE 3]` *A teacher comparing a new piece with earlier work, drafts and a conversation with the student*

**EMMA**  [04:09]
And I mark what?

**GEORGE**  [04:11]
What's theirs. Content the student didn't produce can't earn marks.

**EMMA**  [04:15]
I sign to say it's their own work. How am I supposed to know?

**GEORGE**  [04:20]
The guidance says you must be familiar enough with the candidate's general standard to judge whether the work is within their capabilities. And you mustn't accept work you can't authenticate.


`[CUE 4]` *A detection score of ninety per cent with a label reading a question, not an answer*

**EMMA**  [04:32]
What makes that easier?

**GEORGE**  [04:34]
Seeing drafts and notes along the way. Having some of it done under supervision. And asking them to talk you through it.

**EMMA**  [04:43]
What should make me look twice?

**GEORGE**  [04:45]
A sudden change in style or vocabulary. References you can't find. Content that doesn't match what was taught. And a student who can't explain their own work.


`[CUE 5]` *Two routes from a suspicion: before the declaration to the centre's policy, after it to the awarding organisation*

**EMMA**  [04:56]
I ran it through a detector. Ninety per cent AI.

**GEORGE**  [05:00]
Then you have a question, not an answer. The guidance is careful here. Detection tools vary in accuracy. They produce false positives, and research keeps finding they're more likely to flag people writing in a second language.

**EMMA**  [05:15]
So I can't use the score.

**GEORGE**  [05:17]
You can use it as part of the evidence. Never the whole of it. And never tell a student or a parent that a tool has proved anything.

**EMMA**  [05:28]
So what do I actually do?

**GEORGE**  [05:31]
Investigate. Gather his earlier work, the drafts, any acknowledgement. Talk to him and ask him to explain it. Keep a record.

**EMMA**  [05:39]
And who do I tell?

**GEORGE**  [05:41]
Your exams officer, either way. If it's before he's signed the declaration, the centre deals with it under its own policy. If it comes to light after, the head of centre must report it to the awarding organisation.

### Sources for the on screen credit

- AI Use in Assessments: Your role in protecting the integrity of qualifications (30 April 2025), Joint Council for Qualifications
- Generative artificial intelligence (AI) in education (updated 12 August 2025), Department for Education
- Suspected malpractice: policies and procedures, Joint Council for Qualifications

---

*Copyright WAJD Group. Built by WAJD AI.*