Where does AI appear in everyday life? — recognise the suggestion and decide for yourself

AI often appears as a suggestion

You are writing a message to a friend on your phone. Above the text, the keyboard suggests the next word based on what you have just typed and the words you have used before. You can tap the suggestion or ignore it and type your own word. A small feature like this may use AI, even though it does not look like a robot or a separate chat. Whether it does depends on the keyboard and the feature.

Remember the input–suggestion–decision model from the previous guide? The same model appears here: the keyboard received the text you typed as input, suggested a word, and you decided whether to tap it or ignore it. The same structure appears in many other everyday features, such as the next song on a playlist or the next episode on a streaming service.

Make your first observation in ten minutes

1. 1

   Choose one familiar service

   Think of a music service, photo library, online shop or phone keyboard, for example. Choose a function that suggests something.

2. 2

   Copy the practice request

   The request below asks you which feature you mean if you have not already said. Send it and answer the question.

3. 3

   Look at the suggestion through three questions

   What information does the feature use? What does it produce? Can I accept, change or ignore it?

4. 4

   Test it against your own purpose

   Write one sentence about whether the suggestion matched your goal. If it did not, look in the settings or help pages for a way to influence it.

Think of one feature on your phone or favourite app that suggests something to you — for example, a keyboard word suggestion, the next song on a playlist, the next episode on a streaming service or a map's suggested fastest route. If I have not yet told you which feature I mean, ask me first. Once you know the feature, describe it under three headings: Input (what information the feature probably uses), Suggestion (what it offers me) and Decision (what I can accept, change or ignore myself). Clearly distinguish established information from a general example, and do not claim to know exactly how this particular service works internally.

Check your previous answer. Mark every point that is only a guess with a question mark, then rewrite the rest more concisely using the same three headings.

Use the same feature you just described. Make me a four-point checklist: what information it uses, where I can find the service's own description, which setting or feedback option I can try, and how I can compare the suggestion with my own purpose. Do not guess the names of a particular service's settings.

A suggestion is not a decision

A recommendation can save time, but it is always a suggestion based on some information. When a feature offers text or an image, compare it with what you wanted to achieve.

Return to the feature you have described in this conversation. Turn your observation into one clear question you could put to the service's own help pages or customer support. The question should establish what the suggestion is based on and how I can influence it. Do not name the service or invent a setting name.

What does this feature suggest, what information does it need to work, and where can I find the service's own description of it?

Check this specific point

1. Open the service's own help pages when you want to know what information a function uses.
2. See whether you can ignore, correct or remove the suggestion.

Once you know the basics

Next you will explore an AI agent: an assistant that can do more than suggest things — it can also use tools and take actions with limited permission. You will learn to distinguish a simple suggestion from an action that still needs your approval.