Section
Artificial intelligence in plain words: how to use it well, how to check it, and how not to be fooled by it.
Scam checker. Check a text, email, website or call before you act →
Start here
A start-here guide to AI: what an AI assistant is, what it does well and badly, how to hold a first conversation, how to check its answers, what to keep private, how to use it before a health appointment and how to see through voice clones, deepfakes and fake reviews. It ends with free places to learn in the UK and US and a seven-day plan.
Read the guideHow 'family emergency' calls work now that criminals can clone a voice from a short clip, and what the FTC, FBI, FCC, AARP and UK fraud bodies advise. The defences in order of usefulness: call back, a family safe word, no unusual payments, no secrets, plus where to report in the UK and US.
Where people actually meet AI fakes (celebrity investment videos, romance scams, fake reviews, breaking-news images) and a six-step method for checking them. Includes what UK and US law now says about fake reviews and where to report the scam versions.
How to use an AI assistant to prepare for a medical appointment: a short timeline, three priority questions, plain-word explanations of terms you have already been given, and a write-up afterwards. Plus what the WHO and the AI makers themselves say about its limits, and a clear list of what not to ask it.
What the makers of ChatGPT, Claude, Copilot and Gemini say happens to your chats, and what the UK's NCSC and Canada's privacy commissioner advise. A plain never-list, a fine-to-share list, and how to get help with a private document without handing it over.
Why a fluent answer from an AI assistant can still be wrong, what the makers and independent research say about it, and which answers need checking. A six-step routine and five warning signs for anything touching health, money, law or deadlines.
A first-timer's guide to AI assistants such as ChatGPT, Claude, Copilot and Gemini: what you need to start, how a conversation differs from a search, how to ask so the answer is useful, and five ready-to-type first requests. It also covers what not to type in, why answers need checking, and where to learn alongside other people.
How an amateur codebreaker's list of unsolved ciphers became a test bed for AI, what the machines can and cannot yet do, and how to try codebreaking yourself, by hand and with a free AI assistant you check line by line.
How a suitcase that bumped into a wall became an AI guide for blind travellers, what the trials have shown, and how to try AI that describes what a camera sees, with the two habits that keep it safe.
Margaret Colling, a retired librarian and grandmother of three from Morecambe, joined the People's Panel on AI in 2023. She is now a leading champion of Let's Talk AI, a community-designed awareness campaign from the University of Sheffield that launched on bus stops and social media in April 2026.
