Free AI Starter Guide

A Short History Of AI

AI can feel as if it appeared out of nowhere a couple of years ago. In fact, people have been working on it for more than 70 years. Here's the short version.

The Early Days

In 1950 the British mathematician Alan Turing wrote a famous paper that began with a simple question: "Can machines think?" He suggested a test. If you swapped typed messages with a machine and couldn't tell it apart from a person, perhaps you'd have to say it was thinking.

Six years later, in the summer of 1956, a small group of scientists met at Dartmouth College in New Hampshire, USA, to work on the idea. Their workshop is where the name "artificial intelligence" was first used.

Games were an early favourite, because they have clear rules and a clear winner. Turing wrote one of the first chess programs on paper, before there was a computer that could run it. In the 1950s Arthur Samuel at IBM built a checkers (draughts) program that got better by playing game after game against itself.

Chatbots And Chess Champions

In 1966 Joseph Weizenbaum at MIT built ELIZA, one of the very first chatbots. It worked by turning your own words back into questions. It understood nothing at all, yet some people who tried it felt it was really listening to them.

In 1997 IBM's Deep Blue computer beat the world chess champion, Garry Kasparov, in a six-game match. It made headlines around the world, but Deep Blue could do exactly one thing: play chess.

AI Moves Into Your Pocket

In 2011 Apple put Siri on the iPhone 4S, and talking to your phone started to feel normal. Amazon's Alexa followed in 2014, inside the Echo speaker. Quietly, AI was also turning up in the email spam filter, your bank's fraud alerts, sat-nav routes and photo apps that sort pictures by face. In 2016 a program called AlphaGo, from the Google-owned lab DeepMind, beat one of the world's best players at Go, a board game far harder than chess for a computer.

Why It Suddenly Took Off

On 30 November 2022 a company called OpenAI released ChatGPT. You didn't need to understand anything technical. You just typed a question in plain words and it wrote back like a person. It was free to try, and OpenAI's boss said more than a million people had signed up within five days. Google, Microsoft and others raced to release their own.

So why then, after 70 years? Three things came together:

• Much more computer power. Graphics chips, first made for video games, turned out to be very good at training AI.
• Huge amounts of writing to learn from. Decades of websites, books and articles gave these systems endless examples of how people write.
• A better design. In 2017 Google researchers published a new way of building AI, called the "transformer", that is much better at handling language. It's the T in ChatGPT.

Add a simple chat box that anyone could use, and AI went from the science pages to the kitchen table almost overnight.

Back to contents

What AI Actually Is, And Isn't

AI, short for artificial intelligence, is a broad name for computer programs that do jobs we'd normally expect to need a person: understanding speech, recognising a face, writing a reply. You've probably been using it for years without calling it that, every time your phone guessed the next word of a text.

Chatbots, In Simple Terms

A chatbot is AI you talk to in everyday words, by typing or speaking. ChatGPT, Gemini, Claude and Microsoft Copilot are the best known. Under the hood (or the bonnet, if you're in the UK) is something experts call a "large language model". That sounds grand, but the idea is simple. It has read an enormous amount of writing, and it has become very good at predicting which words should come next. When you ask it something, it builds an answer one word at a time.

A good way to think of it: a very well-read helper who has skimmed almost everything, remembers most of it a little hazily, and never gets tired of your questions.

What It's Good At

• Explaining something confusing in simple terms, as many times as you need.
• Drafting and tidying up writing: emails, letters, a speech, a card message.
• Coming up with ideas, from gift lists to what to cook tonight.
• Boiling a long document down to the main points.
• Answering follow-up questions patiently, without ever sighing at you.

What It Isn't

• It isn't a person. It can sound warm and friendly, but it has no feelings and doesn't know you unless you tell it things.
• It isn't always up to date. Some chatbots can search the web and show you links. If yours does, the links are worth clicking.
• It isn't your doctor, lawyer or financial adviser. It can help you understand things and prepare questions. Decisions belong with you and a qualified professional.

Why It Sometimes Gets Things Confidently Wrong

Because a chatbot predicts likely-sounding words rather than checking facts, it can give an answer that sounds completely sure of itself and is simply wrong: a made-up date, a book that doesn't exist, the wrong opening hours. Experts call this "hallucinating". It isn't lying. It just can't always tell the difference.

The companies say so themselves, in the small print under the chat box. So use it as a starting point, not the final word. Ask it "How sure are you?" or "Where could I check that?", and for anything to do with health, money or the law, check with a professional or an official website.

Back to contents

Getting Started Safely

You only need one. All three below are free to start, and any of them is a fine first choice. Pick one, get comfortable, and only try another later if you're curious.

Setting Up An Account

1. Go to the real thing. Type the address above into your web browser yourself, or install the app from the Apple App Store or Google Play. Check the app is made by OpenAI, Google or Anthropic. Copycat apps exist.
2. Choose "Sign up". You can use your email address, or an account you already have, such as Google or Apple, if it's offered.
3. Look for a code. You'll usually get an email with a code or link to confirm it's you. You may also be asked for your mobile number to get a text code. That's normal.
4. Lock it. Use a password you don't use anywhere else, and turn on two-step verification in the settings if it's offered.
5. Don't pay to start. The free versions are genuinely free. If something asks for card details just to try it, stop and check you're on the real website.

ChatGPT will even let you try a few questions without making an account at all, which is a gentle way to dip a toe in. Wondering about paid plans? See what they cost in the USA and UK.

Finding Your Way Around The Screen

Whichever you pick, the screen looks much the same. There's a box at the bottom where you type, your conversation in the middle, and a menu holding your past chats. Here's what the main buttons do.

Start a new chat whenever you change subject. Within one chat it remembers what you've already said, so a fresh chat for each topic keeps things tidy.

Back to contents

How To Ask Good Questions

There are no magic words, and you can't break it by asking the "wrong" thing. The best advice is simple: talk to it the way you'd talk to a helpful, patient person behind a counter.

Give It Some Context

The more it knows about what you want, the better the answer. A few words about who it's for, what happened, how long it should be and the tone you'd like make a big difference.

Keep The Conversation Going

Treat the first answer as a first draft. You can reply just as you would to a person:

• "Can you make that shorter?"
• "Make it sound a bit friendlier."
• "Give me three different options."
• "I don't understand the part about interest rates. Can you go over it again?"
• "What questions should I be asking about this?"

"Explain It More Simply" Is Always Allowed

Never feel daft asking. It doesn't judge, and it won't get bored. Try "Explain it more simply", "Use an everyday example" or "Put that in five short bullet points". You can also ask it to use American or British spelling, or to avoid jargon altogether.

You Can Talk Instead Of Type

Tap the microphone and just speak. It's often quicker, and it's a big help if small keyboards are a struggle.

Back to contents

Everyday Wins

This is where AI earns its keep. Here are some of the most useful everyday jobs to try first, with a question you can copy.

Back to contents

Staying Safe And Private

A few simple habits keep you safe, both when you use AI and when scammers try to use it on you.

What Never To Share

What you type is stored on the company's computers. Depending on your settings it may be used to improve the AI, and some conversations may be reviewed by staff. Look in the settings for "data controls" or "privacy" to see your choices. A good rule of thumb: treat the chat box like a postcard, not a sealed letter.

More detail on settings: Is ChatGPT safe?

How AI-Powered Scams Work

Scammers use the same tools as everyone else. The old giveaways, like bad spelling and clumsy English, are disappearing. Watch out for:

• Polished scam emails and texts that look just like your bank, a delivery firm or the tax office.
• Voice clones. The US Federal Trade Commission (FTC) warns that a scammer needs only a short clip of someone's voice, perhaps from a video posted online, to copy it. Then they call you, sounding like your grandson or daughter, in trouble and needing money fast.
• Fake videos of famous people recommending an investment, a miracle cure or a "government scheme".
• Fake chatbot apps and websites that charge for something that's free, or steal your login.
• Online romances with someone whose photos, messages and even video calls are made by AI.

Agree a family safe word face to face, not by text or email, and pick something nobody could guess from social media. More on this in our guide to spotting a voice-cloning scam.

How To Spot Fake Images And Video

Honestly, it's getting harder, and there's no single giveaway. These clues still help:

• Look closely at hands, teeth, ears, jewellery and writing in the background. They can be blurred, jumbled or just a bit odd.
• Watch the mouth. In fake videos the lips may not quite match the words, and the voice can sound flat or oddly smooth.
• Check where it came from. Who posted it first? Are trusted news organisations reporting the same thing?
• Search the picture. A reverse image search, such as Google Lens, shows where else a photo appears online.
• Notice how it makes you feel. Fakes are often designed to make you angry, scared or excited enough to act before you think.

A famous face selling an investment or a cure is a scam until proven otherwise. Real celebrities don't send you private messages asking for money.

If Something Feels Wrong

Stop. Don't pay, don't click, and talk to someone you trust. Then report it. You won't be wasting anyone's time.

Not sure about a message right now? Try our Is this a scam? checklist.

Back to contents

Checking What It Tells You

Chatbots are clever, but they aren't always right. The good news is that checking an answer is usually quick, and it soon becomes a habit. Think of it like advice from a well-read friend: worth listening to, but you'd still check the train times before you set off.

Why Checking Matters

As we saw earlier, a chatbot writes what sounds likely, not what it has checked. Most of the time that works well. But it can mix up dates, invent a quote, give an out-of-date price or recommend a book that doesn't exist, all in the same confident tone. It won't always warn you when it's guessing.

Five Quick Ways To Check An Answer

1. Ask where it came from. Try "Where could I check that?" or "Please give me links to your sources." Then click the links and make sure they really say what it claimed. If a link doesn't work, or says something different, trust the link, not the chatbot.
2. Check one trusted source yourself. For benefits, tax, pensions or the law, go to the official government website. In the USA these usually end in .gov. In the UK, look for GOV.UK. For opening times and prices, go to the company's own website.
3. Ask again, a different way. Reword the question, or start a new chat and ask again. If you get two different answers, that's your cue to check further.
4. Look hard at the details. Exact numbers, dates, names, quotes, phone numbers and web addresses are where mistakes most often slip in.
5. Ask it to argue the other side. "What might be wrong with this answer?" or "What would someone who disagrees say?" often brings up things it skipped.

When You Can Trust It More

Not everything needs checking. A rough guide:

• Fine as it is: ideas, first drafts, recipes, gift lists, rewording your own letter, explaining a word. If it's a little off, there's no harm done, and you'll usually spot it.
• Check it first: facts you'll rely on or pass on to others, such as dates, prices, opening times, travel details and rules.
• Ask a real expert: anything about your health, your money or the law, where a wrong answer could cost you.

Health, Money And Legal Matters

A chatbot can be a real help here, as long as it stays in the passenger seat. Use it to understand the words in a letter, get your thoughts in order, or write down questions to take to an appointment. Then let a qualified person make the call.

• Health. Never start, stop or change a medicine because a chatbot said so. Ask your doctor or pharmacist first.
• Money. Be very wary of investment tips from a chatbot or a social media post. Before you trust an adviser or firm, check they're registered (see below).
• Legal. Laws differ between US states, and between England, Scotland, Wales and Northern Ireland, and chatbots often mix them up. For anything serious, such as a will, a dispute or a contract, speak to a lawyer (USA) or solicitor (UK).

A handy habit: end important questions with "How sure are you, and what should I double-check?" It will often tell you which parts are shaky.

Back to contents

Going Further

Once you're comfortable typing questions, there's a lot more AI can do. None of it is compulsory. Pick what sounds useful and leave the rest for another day.

Talking To A Voice Assistant

Voice assistants let you ask out loud and hear the answer, which is handy when your hands are busy or your eyes are tired. You may already have one:

• Siri on iPhones, iPads and Macs. Say "Siri" or "Hey Siri", or hold the side button.
• Gemini on Android phones, where it has now taken over from Google Assistant. Saying "Hey Google" or holding the power button usually starts it.
• Alexa on Amazon Echo speakers. Amazon is rolling out a chattier version, called Alexa+, in the USA and UK.
• Chatbot apps such as ChatGPT, Gemini and Claude have a voice mode too. Tap the voice or microphone button and simply have a conversation.

Good first things to ask out loud: "Set a timer for 20 minutes", "Remind me to take my tablets at 8pm", "What's the weather tomorrow?" or "Call my daughter".

Smart speakers listen out for their wake word, so it's worth a look at the privacy settings in the app. Most let you review and delete recordings, and have a button to switch the microphone off.

Photo Tools

Your phone's camera can now do much more than take pictures. Point it at something and ask about it:

• Identify things. Google Lens can tell you what a plant, landmark or object is, and translate a menu or sign abroad. PictureThis is made for plants, and Merlin Bird ID can name a bird from a photo or its song.
• Read and describe. Be My Eyes and Microsoft's Seeing AI can read out letters and labels and describe what's in front of you. A real help if your eyesight isn't what it was.
• Ask a chatbot about a photo. Tap the plus button in ChatGPT, Gemini or Claude, add a photo, and ask something like "What's this warning light on my dashboard?" or "What could I cook with what's in this fridge?"
• Find and tidy your photos. Google Photos and Apple's Photos app can search your pictures by what's in them, such as "beach", "dog" or "birthday cake", and many phones can now remove an unwanted object from a picture.

A word of caution: photos you give an AI app are sent to the company. Keep out pictures of documents, bank cards and anyone who hasn't agreed to it.

AI Already On Your Phone

• Captions. Live Caption on Android phones and Live Captions on iPhone can show words on screen for videos and calls. Google's Live Transcribe turns the conversation around you into text.
• Writing help. Newer iPhones and many Android phones can tidy up or summarise a message, an email or a long web page for you.
• Screening unknown callers. Some phones can answer a call from an unknown number for you and show who it is and what they want before you pick up.
• Spam and scam warnings. Your phone or network may flag suspected spam calls and scam texts. Treat the warning as a helpful hint, not a guarantee.

Many of these features live in your phone's Settings, under Accessibility. They're built in and free.

Your 30-Day Plan

Ten minutes a day is plenty. Skip a day, repeat a day, go at your own pace. Here's a gentle plan to build your confidence:

1. Week 1 (days 1 to 7): Say Hello. Pick one chatbot and set up your account safely, as in "Getting Started Safely". Then ask one question a day: something you've always wondered about, a word you don't know, or what to have for dinner.
2. Week 2 (days 8 to 14): Make It Useful. Draft a real email or letter, make sense of a bill or leaflet (with your details covered), plan a day out, and try a recipe from what's in the cupboard. Practise follow-ups: "shorter", "friendlier", "explain it more simply".
3. Week 3 (days 15 to 21): Talk And Snap. Try voice mode, or ask your phone's voice assistant to set a reminder. Use your camera to identify a plant or translate a label. Turn on captions if they help.
4. Week 4 (days 22 to 28): Check And Stay Safe. Fact-check one answer using the five steps in "Checking What It Tells You". Look through your chatbot's privacy settings. Agree a family safe word face to face, and try our scam checklist.
5. Days 29 and 30: Look Back And Choose. Which three uses helped you most? Keep those. Then show someone else what you've learned. Teaching it is the best way to remember it.

Back to contents

The Future Of AI

You'll hear a lot of big claims about what AI will do next. Some will come true, some won't, and nobody knows the timetable for sure. So instead of guessing, here's what is already happening or being properly tested, and what it could mean for you.

Health

Some of the biggest changes are quiet ones. AI already helps specialists read some scans and medical images, for example looking for signs of eye disease or skin cancer. In the USA, the Food and Drug Administration (FDA) keeps a public list of the AI-enabled medical devices it has authorised. In the UK, the NHS is running a large trial, called EDITH, to test whether AI can help experts spot breast cancer in screening mammograms. In both cases, trained people still make the final decisions.

You may also come across AI behind the scenes at appointments, such as software that writes up notes so the doctor can look at you rather than the screen. You can always ask how it's being used, and say if you're not comfortable with it.

Home

Voice assistants are getting chattier and better at following a conversation. Amazon's Alexa+ is rolling out in the USA and UK, and Google's Gemini has taken over from Google Assistant on Android phones. Alongside reminders and timers, features like fall detection on some smartwatches can help people live independently for longer, and give family some peace of mind.

The trade-off is privacy. More helpful usually means the device knows more about you. Read the settings, and only share what you're comfortable with.

Transport

Many newer cars already have driver aids, such as automatic emergency braking and help staying in your lane. In several US cities, including Phoenix, San Francisco and Los Angeles, you can now book a Waymo taxi with nobody in the driver's seat. In the UK, Uber and Wayve began offering self-driving rides in London in September 2026, with a trained safety driver on board. Fully driverless taxis there still need further approval.

For people who no longer drive, or would rather not, this could one day mean more freedom to get around. For now it's limited to certain cities, and it's perfectly sensible to wait and see.

Work

AI is now built into everyday office software, such as Microsoft Copilot in Word and Outlook, and Gemini in Google's Gmail and Docs. Companies are also testing "AI agents" that carry out tasks step by step, such as filling in forms or booking things. Some jobs will change and new ones will appear, but experts honestly disagree about how much and how fast.

If you're still working, or thinking about part-time work or volunteering, being comfortable with these tools is a genuine plus. And your experience and judgement, knowing when an answer doesn't feel right, are exactly what AI lacks.

What It Means For You

• More help with everyday life. Captions, reading aids, reminders and voice control are already making life easier for people with hearing, sight or mobility challenges.
• Smarter scams, too. The same tools make fake voices, videos and messages more convincing, so the habits in "Staying Safe And Private" matter more than ever.
• You stay in charge. You don't have to use every new gadget. Try what's useful, ignore what isn't, and ask how your information will be used.
• It's never too late to learn. The basics in this guide will still serve you well as the tools change.

Be wary of anyone who says they know exactly what AI will do in five or ten years, especially if they're selling something. A good question to ask: "What's working today, and how do you know?"

Back to contents

AI Jargon Made Simple

AI comes with a lot of jargon. Here are the words you're most likely to come across, in simple terms. You'll find more in our online glossary.

AI agent

An AI that can carry out a task for you, step by step, such as filling in a form or comparing prices. Always check what it's about to do before it spends money or sends anything.

Agentic AI

AI that can take actions for you, not just answer. Another way of saying "AI agents".

Algorithm

A set of step-by-step instructions a computer follows, a bit like a recipe.

App

A program on your phone, tablet or computer.

Artificial intelligence (AI)

Computer programs that do jobs we'd normally expect to need a person, like writing, understanding speech or recognising a face.

Autonomous

Able to carry on by itself without asking you at every step. Keep an eye on any AI that works this way.

Bias

When an AI's answers lean unfairly one way, because of the examples it learned from.

Chatbot

A program you chat with by typing or talking, such as ChatGPT, Gemini, Claude or Microsoft Copilot.

Cloud

Computers run by a company and reached over the internet. "Saved in the cloud" means stored on their computers, not just on your device.

Computer use

When an AI moves the mouse, clicks and types on a screen the way you would. Sometimes called an "operator".

Deepfake

A fake photo, video or voice made by AI to look or sound like a real person.

Generative AI

AI that creates something new, such as text, pictures, music or video, from your description.

GPT

Short for "generative pre-trained transformer", the technology behind ChatGPT. The "transformer" is the 2017 design mentioned in the history section.

Hallucination

When an AI confidently states something that isn't true. It's a mistake, not a lie, so check anything important.

Large language model (LLM)

The "engine" inside a chatbot. It learned from huge amounts of writing and predicts which words should come next.

Phishing

Fake emails, texts or calls that try to trick you into giving away passwords, codes or money.

Prompt

The question or instruction you type or say to an AI.

Smart speaker

A speaker with a voice assistant built in, such as an Amazon Echo or Google Nest.

Training data

The examples an AI learned from, such as books, websites and pictures.

Two-step verification

A second check when you log in, usually a code sent to your phone, so a stolen password alone isn't enough.

Voice assistant

An AI you talk to out loud, such as Siri, Alexa or Gemini.

Voice cloning

Copying someone's voice with AI. Scammers use it to pretend to be family, which is why a family safe word helps.

Heard a word that isn't here? Ask a chatbot: "Explain [the word] in simple terms, with an everyday example." It's very good at exactly that.

Back to contents