Everyday life · England
Using artificial intelligence
Your browser or device provides the reading voice.
This guide is for people in England.
You can read it with someone you trust.
AI is a type of computer tool.
Some tools answer questions or help with writing.
You can choose whether to use them.
Start with one small task.
Ask for short sentences and one step at a time.
AI answers can be wrong.
Even an answer that sounds sure can be wrong.
Check important information on the original website.
Keep passwords and bank details private.
Do not upload private records about yourself or other people.
Some tools cost money.
Check the price and how to cancel.
You decide what you want to say.
Read the message before you send it.
Talk to a person about important health or legal decisions.
Use urgent services if you need urgent help.
What you can do next
You can ask someone you trust for help.
Choose the next step that works for you.
Read about AI safety Web pageThis link opens another website.
It may use harder words.
More useful links
- AbilityNet: background for service providers Web page
Provider-focused information about accessible AI and human support.
- NCSC: AI safety Web page
Understand inaccurate answers and other risks.
These links open other websites.
They may use harder words.
Using artificial intelligence
Try useful AI tools, protect private information and check important answers.
Adults, young people and supporters · England
What AI means
Artificial intelligence, or AI, is a name for computer systems that perform tasks such as recognising patterns or generating text. A chatbot is a tool you can type or speak to. It produces answers using patterns it has learned. A friendly voice or confident answer does not mean it understands you as a person would. Different tools have different abilities and limits.
Try one useful task
You could try asking for ideas for a hobby, a shopping-list layout, a practice conversation or a simple explanation of a public document. Ask for one step at a time and words you understand. For example: “Help me write a polite message asking a venue about step-free access. Use short sentences. Leave spaces for my details.” Read and change the result before sending it. You stay in charge of what it says.
Check important answers
AI can invent facts, quotations and website links. It can repeat unfair assumptions about disabled people. Check the original source yourself: does the page exist, say what the AI claims, apply to England and give current information? Use the relevant NHS, GOV.UK or service page. Asking the same chatbot whether it is right is not an independent check. For health, benefits, money or legal decisions, get suitable human advice as well.
Protect your information
Do not paste passwords, bank details, identifying health records or private information about other people into a public chatbot. Use made-up names and remove identifying details when trying an example. Removing a name alone may not make a document anonymous. Read the service’s privacy settings and terms: storage, sharing and training arrangements vary. A supporter should explain what would be uploaded and respect the person’s choice. Work records need the organisation’s authorised process.
Choose an accessible tool
Try the tool with a small, low-risk task before relying on it. Check whether you can read the answer, use your communication method, correct a mistake and stop it. Voice input, reading aloud and other accessibility features vary. Check age rules, subscription costs, renewal terms and cancellation before paying. A service should explain how to reach a person when its chatbot cannot help. Ask for a reasonable adjustment where a disability creates a barrier.
Stay in control
AI cannot give reliable personal permission to drive, change medicines, ignore a legal deadline or spend money you cannot afford. Do not let it send messages or buy things without understanding and agreeing to the action. If an answer frightens, pressures or confuses you, stop and talk to someone you trust. For urgent health or safety problems, use the emergency and professional support routes on this site.
Check unexpected requests
Computer-generated voices, pictures and messages can be used in scams. If someone unexpectedly asks for money or personal details, pause. Contact the person or organisation using details you already trust. Do not use a new number or link supplied in the suspicious message. Our scams guide explains reporting and support.
Words you could use
“Please help me try this tool, understand its privacy settings and check the answer before I use it.”
More useful links
- AbilityNet: background for service providers Web page
Provider-focused information about accessible AI and human support.
- NCSC: AI safety Web page
Understand inaccurate answers and other risks.
These links open other websites.
They may use harder words.
Where the information comes from
Links checked 1 November 2026.
These links open other websites.
They may use harder words.
- NCSC: AI and cyber security Web page
AI errors, bias and security risks.
- NCSC: public chatbot information risks Web page
Why sensitive information needs protection; individual providers’ current terms must be checked.
- ICO: accuracy of generative AI outputs Web page
Potential harms when generated information is wrongly treated as fact.
- AbilityNet: AI and accessibility Web page
Accessibility barriers, error correction and access to a person.
A page for you to fill in
You can type your answers and download a copy.
You can also print a page to write or draw on.
Someone you trust can help you.