What is AI?
AI isn't new
Artificial intelligence (AI) is the use of computer systems to solve problems and make decisions. It’s already a part of everyday life – you’ve probably come across it in the form of personalised suggestions on social media, shopping sites or route-planning apps.
However, the technology is developing rapidly and throwing up many new challenges for schools.
What is generative AI?
Generative AI takes a written prompt and runs it through an algorithm to generate new, ‘natural’-seeming content – i.e. that seems like it was created by a person, not a computer. Tools include:
- Chatbots such as ChatGPT, Google Gemini and GrammarlyGO, which generate text
- Text-to-image programs like DALL-E and Midjourney, which create images (some programs can make AI-generated animations and near-photorealistic videos, too)
How to respond to AI
Update your policies
You may decide to set out your overall approach in an artificial intelligence (AI) policy. This should include how you decide which AI tools are safe for different tasks.
If you haven't already, you should also review your:
- Data protection policy: this demonstrates how you will protect personal data and comply with the UK GDPR
- Homework policy: this sets out the acceptable use of AI by pupils. You can use and adapt our sample AI clause text
- Exams policy: the Joint Council for Qualifications (JCQ) classifies AI misuse – where a pupil submits AI-written work as their own – as malpractice. Check whether your policies that cover exams, assessment, coursework and plagiarism reflect this, and revise them if needed
You'll need to review all of these policies regularly, or whenever there is a significant change in how you use AI.
Explain your rules on AI use to pupils
Make sure pupils know that using AI without crediting it, isn't allowed in exams, coursework or any work that’s internally assessed to count towards a qualification. Remind pupils of this when they have exams and coursework coming up.
If you’re extending this policy to homework and other independent study, communicate this clearly to all pupils.
Have an open dialogue with pupils about how and when AI tools can be used to support learning, and when they shouldn’t be.
Help keep pupils safe online
AI tools can pose new and unique safeguarding risks. This includes pupils having harmful online interactions with generative AI tools, such as where AI chatbots may validate thoughts around self-harm and suicide. AI can also be used to digitally alter or wholly generate sexual images of children, known as 'deepfakes' or 'deep nudes’.
Make sure:
- Your filtering and monitoring systems are effective at preventing inappropriate AI-generated content or harmful AI interactions in school
- Pupils are taught about the specific risks of AI as part of your school's approach to online safety
Don't enter sensitive information into an AI tool
Be aware that any text entered into an AI tool is potentially being made public. If you decide it's strictly necessary to use personal data in generative AI tools, you must make sure that all steps are taken to protect the data, and that the products and procedures comply with:
- Data protection legislation
- Your data privacy policies
Read more about the dos and don'ts for data protection when using AI.
You may also be targeted by fraudulent emails, which are often AI-generated and very convincing. Use our data protection training resources to keep your staff vigilant to cyber threats – encourage your staff to take our UK GDPR and cyber security eLearning courses.
Protect staff and pupils' intellectual property
Intellectual property is protected by copyright law, which is different to data protection law. It includes any original work created by a person, such as:
- Essays, homework or any other materials written or drawn by a pupil (not multiple-choice questions)
- Lesson plans created by a teacher
- Prompts entered into generative AI tools
You must have the consent of either pupils or staff (unless an exception to copyright applies) to enter their intellectual property into any AI tool that will store or collect this information for commercial purposes, such as training or improving the AI (e.g. many free-to-use AI tools). If pupils are too young to give consent, you should get permission from their parents/carers.
Alternatively, choose a tool that doesn't store this information, or gives you the option to opt out of data collection.
Be aware of 'secondary infringement' of copyright
This happens when an AI is trained on material it doesn't have permission to use, which you end up unknowingly using through the AI tool. The DfE provides 2 examples of this:
- Publishing a policy created by an AI tool that used input taken from another school's policy without that school's permission
- Using an image on a website created by an AI tool using input taken from the copyright holder without their permission
This is explained in the DfE's policy paper on generative AI in education.
AI could save you time …
You can use AI to cut down on some of your workload. For example, it could help you:
- Create a comment bank to use when report writing
- Come up with ideas for charity fundraising activities
- Write quiz questions to check pupils’ knowledge
… but it’s not always reliable …
AI tools are only as accurate as the information they’re trained on. They may generate responses that are incorrect, biased or inappropriate.
Many tools are based on a defined set of information, so won’t be able to accurately give you answers about anything that has changed after data was inputted – e.g. new statutory policy requirements or current events.
… so it's important to check AI-generated results carefully
Use AI tools as a starting point, but always check and adapt the results so they're:
- Taking the best interests of staff, pupils and the school into account
- In line with your school/trust policies, procedures and guidelines that cover generative AI
Ofsted will record your use of AI
Ofsted won't ask if you use AI or actively search for your use of AI, however, if inspectors come across instances where your school is using AI, or if you have an explicit policy on pupils using AI, they might consider the impact that the use of AI has on the outcomes and experiences of pupils. This includes how your school uses AI, but also how you respond to the use of AI by others, such as parents/carers and pupils.
Where inspectors see your school uses AI or has a policy on its use, inspectors will record:
- How your leaders have chosen to implement it
- Its impact
- How it is monitored
- Your checks and balances to make sure it is accurate, safe and used in the best interests of pupils
Ofsted will assess the risks of AI when evaluating other areas
Inspectors will evaluate your management of AI-related safeguarding risks under the safeguarding judgement area, and not in isolation. Not all uses of AI will be relevant to an inspection, either.
Inspectors may ask what steps you've taken to make sure your school's use of AI takes risks into consideration, such as around data protection, safeguarding, and bias and discrimination.
This is explained in Ofsted's guidance on how it looks at AI during inspection and its operating guide for inspectors. You can also read our more detailed summary of how Ofsted inspects artificial intelligence (AI).
Share this information with your staff and governors/trustees
Download this article as a document that you can adapt to reflect your context and approach to AI.
Share it with your staff and governors or trustees, to make sure everyone’s confident on the basics of AI.
Try out The Key's AI-powered assistant
Visit KeyGPT and:
- Give it tasks on your to-do list. It's a customised version of ChatGPT, personalised to your setting
- Use our task-based AI Wizards to create bespoke letters or generate personalised interview questions, job descriptions and adverts in seconds
Key GPT is designed specifically for school leaders, and it's a safe space for you to experiment with generative-AI.