The crew's sorting model works on bright photos and dark photos now.
"Let's turn it into a helper for the whole station!" Rocket says.
Raven opens her notebook. "What do you notice? We need everyone to trust it."
Nova glows. "What would make you trust a helper?" she asks. "Make a list."
A science agency that sets standards lists qualities good AI should have.
When it has them, people can trust it more.
| Quality | In kid words |
|---|---|
| reliable | It works well again and again. |
| safe | It does not cause harm. |
| secure | It is protected from people who try to break in. |
| fair | It works well for everyone. |
| private | It protects people's information. |
| understandable | People can see how it makes choices. |
| Statement | True or false? |
|---|---|
| Trustworthy AI should be fair. | ? |
| Trustworthy AI should protect privacy. | ? |
| Trustworthy AI only needs to work for its builder. | ? |
| People should be able to understand how good AI makes choices. | ? |
Testing a model shows if it is reliable. You did that in Week 7.
Filling data gaps helps make it fair. That was Week 8.
Keeping information safe protects privacy. That was Week 11.
Great start, explorer! Tomorrow we connect these qualities to everything we learned.