The big questions in one place

What is it?

Short answers to the ten questions about AI that come up most often: does it think, is it dangerous, will it replace search engines and when does the EU AI Act apply? Each answer leads on to a more detailed explainer.

An everyday example

Before an argument starts over whether “AI already thinks”, it is enough to read one paragraph and open the linked explainer.

Why does it matter?

Shared starting answers make it easier to talk about AI at home, at school and at work.

A common misconception

That a short answer tells you everything. It is an entry point; the details and sources are in the individual explainers.

In more detail

What is AI?

A broad field of methods and systems that use data or rules to produce predictions, content and decisions.

Does AI think?

Some systems solve tasks that require reasoning, but that does not prove they think the way people do.

Does AI always use the internet?

No. That depends on the app, its permissions and the infrastructure it runs on.

Are all models equally good?

No. They differ by task, language, safety, cost and speed.

Is AI dangerous?

It can cause harm, but the risk depends on the system and how it is used. That is why checking and oversight matter.

Does it learn everything I tell it?

Not necessarily. The current context, stored data and permanent extra training are not the same thing.

Will it replace search engines?

It can help explain things, but to check important facts, you should still open the original sources.

Is AI fair?

Not automatically. Data and design can cause different errors for different groups.

When does the AI Act apply?

Some of the rules already apply; the exact dates and obligations depend on the type of system and its intended use.

What is the first safety rule?

Do not hand over data you have no clear right or need to share, and do not leave important decisions to an unchecked system.

Remember: A good first step is to ask: “How would we check this?”