AI School › The big questions · Lesson 117 of 130

📚 Who decides what AI learns

Behind every AI there are people choosing the data and rules: it is never neutral.

An AI learns from data, that is mountains of texts, images and sounds that someone decided to give it. Who chooses that data? People: engineers, companies, sometimes governments. If you give it only cookbooks, it will get good at recipes and know nothing about history. If the texts it receives speak badly of a group of people, the AI may repeat that prejudice without noticing. So human choices, made at the start, stay inside the machine forever. There is no «neutral AI fallen from the sky»: there is always someone who decided.

Then there are the rules: what the AI can or cannot say. These too are written by people, and different people make different choices. A company may decide the AI should not answer dangerous questions; a government may want it to avoid certain topics. These are important decisions because they touch all of us who use AI. That is why it is fair to ask: «Who built this AI? With what data? With what rules?». Knowing there are human choices behind it makes you a sharp user, able to use the tool without being blindly commanded by it.

🎒 Try it in class (or at home)

Imagine training an AI for your school: which five things would you teach it and which would you avoid? Justify each choice.

← Lesson 116: Can AI make mistakes on purpose? · Lesson 118: Will AI replace humans? →

📱 Download the Android app (beta) iOS coming soon

Say what you mean. Get what you need.

Grace Certified, the AI coach that trains and certifies your prompt engineering, by Agora Intelligence.