G2GGet to grips with tech

Home / Learn

What is an AI chatbot, and is it safe to use?

An AI chatbot is a computer program you can type to as if it were a person, and it writes back. It is handy for explaining things and drafting messages. It is safe to use if you double-check anything important it tells you and never type in private details like passwords or bank information.

Beginner3 min readUpdated

What it is

You may have heard of ChatGPT, Google Gemini or Microsoft Copilot. They are all AI chatbots. You type a question or request in everyday words, and it types an answer back in a few seconds.

“AI” stands for artificial intelligence. The chatbot has learned from a huge amount of writing, so it is good at producing sentences that sound right. That is not the same as always being right.

What it is good for

Google says you can use its Gemini chatbot to:

  • Brainstorm ideas or make a plan.
  • Get a quick, easy-to-understand summary of a complicated topic.
  • Write first drafts of emails, letters and more.

For example, you could ask it to explain a confusing letter in simpler words, or to help you write a polite email to complain about a late delivery.

What it can get wrong

Google is clear that its chatbot can make mistakes. It can present wrong information as if it were fact, and it can be wrong about people too.

  • Double-check anything that matters, like dates, prices, phone numbers or opening times, on an official website.
  • Don’t rely on it for professional advice. For health, money or legal questions, speak to your GP, bank or a solicitor.
  • If it gives you a phone number or a link, find the number yourself on the company’s official website instead.

What never to type in

Google says some chats are read by trained human reviewers, and asks people not to enter anything confidential they wouldn’t want a reviewer to see. Treat every chatbot the same way. Never type in:

  • Passwords, PINs or codes sent to your phone.
  • Bank or card details.
  • Your National Insurance or Social Security number, or medical records.
  • Private details about other people.

A chatbot will never need your password. If a website or message that looks like a chatbot asks for one, close it. It may be a scam.

Privacy settings worth knowing

In Google Gemini (checked 11 Oct 2026):

  • Your chats are kept for 18 months by default. You can change that to 3 months, 36 months or forever, and you can delete chats at any time.
  • Turning off Keep activity stops your future chats being reviewed to improve Google’s services.
  • A temporary chat isn’t used to train Google’s AI. Google still keeps it for 72 hours.

Other chatbots have their own settings. Before you use one, look in its settings for options about chat history and how your chats are used.

Next: read our guide to ChatGPT, try asking an AI chatbot a question on your phone, or learn how scammers use AI to fake a voice.

Sources