Ten questions worth asking first

What is it?

Ten short answers to the questions people ask most often about AI: what it does, why it gets things wrong, what happens to our data and to jobs. No technical vocabulary, made for a first read.

An everyday example

Before you trust a chatbot with a tax calculation or medical advice, ask three questions from this list: where does its information come from, can I open the source myself, and who is responsible if it is wrong?

Why does it matter?

The right questions are the best protection against both too much trust and needless fear. The answers help with everyday decisions, from using tools at work to protecting your privacy.

A common misconception

That a fluent, confident answer is a correct one. A language model can state a false claim just as convincingly as a true one.

In more detail

What does AI actually do?

It takes input and produces output: a prediction, a classification, a piece of text, an image, a recommendation or an action.

How does it learn to do that?

In many modern systems, patterns are formed by learning from data; others use rules, search or a mix of methods.

Why does it get things wrong?

Because the data is incomplete, the task is unclear, or the system produces an answer that sounds convincing but is wrong.

Does AI know what is true?

Not necessarily. A model can state a fact correctly without having checked it independently.

What does it mean when a model is “large”?

Usually that it has many parameters, was trained on a lot of data or uses a lot of computing power. A bigger model is not always the best one for every task.

Am I in control?

That depends on the app: its privacy settings, which tools it is allowed to use, and whether you have to approve its actions.

Can it use my data?

That depends on the provider, the type of account, your settings, the contract and the law. Take extra care with confidential information.

Can I use AI to check information?

Yes, as a helper. But for important claims, ask for primary sources and open them yourself.

Will it replace my job?

It can change tasks, professions and the way we work. The technology alone cannot tell us for certain what will happen to any particular job.

What should I learn?

To understand the task, ask good questions, check the result, protect your data and know when a human has to make the decision.

Remember: Speed and fluent language are not proof that something is true.