Course 8, lesson 80 of 100, Ages 11+
Rules for AI
Keeping AI safe and helpful
Like I’m 5
Games have rules so everyone plays fairly and safely. AI needs rules too, made by people, so it helps and doesn’t hurt.
The big idea
Rules for AI are being written around the world. Some laws ban the most dangerous uses, require testing for high-risk systems like those used in hiring or healthcare, and demand labels on AI-made content.
Rules also come from companies, schools and families. Good rules protect safety and rights while still allowing helpful uses, and everyone, including young people, can have a say in what they should be.
Examples
- EU AI Act: Sorts AI systems by risk level, with the strictest rules for high-risk uses.
- Labels: Some places require AI-made images or chatbots to be clearly labelled.
- School rules: Guidelines on when students may use AI for homework.
How it works
- Countries are writing laws about AI, like rules for risky uses and for labelling AI-made content.
- AI companies test their systems for safety and work to make them follow human values.
- You can help too: report harmful AI content, and ask whether a use of AI is fair.
Check your understanding
- Why does AI need rules?
- Options: So it’s fair and safe for everyone; So it runs faster; So it looks nicer.
Answer: So it’s fair and safe for everyone. Rules protect people and make sure AI is used responsibly. - What do many AI laws require for high-risk systems?
- Options: Careful testing and safeguards; No rules at all; That they're painted red.
Answer: Careful testing and safeguards. High-risk uses like hiring or healthcare need extra checks.
Remember
Good rules and careful testing keep AI safe.
Talk about it
If you could make one rule for AI, what would it be?
Go deeper
Examples include the EU AI Act, which sorts uses by level of risk, and guidelines in many other countries. Researchers work on alignment (making AI pursue the goals we intend), interpretability (understanding what models do inside) and testing for dangerous capabilities.