Chapter 11 / 17
What does an AI expert do?
AI experts do more than write code. They ask questions, work with others, test ideas, and learn from mistakes.
Understand the problem
Before building a tool, an expert talks with the people who might use it. A clear problem helps a team choose whether AI is even needed.
Build, test, improve
Experts examine data, try a design, and check results for errors or unfairness. Feedback from users helps them improve or change course.
Ava's team tests an idea
Ava's team wanted to make a tool that could help students find books. The coder built a search feature, the librarian checked the book details, and the students tried different questions. One search kept showing cookbooks when they asked for stories about space.
The team listened, changed the way books were described, and tested again. Ava saw that AI work is teamwork: people understand the problem, build, test, and improve. Finding a mistake was useful because it showed what to fix.
A real-life example
A team building video captions might include programmers, people who know about language, and viewers who use captions. Each person can notice different problems.
Plan a helpful tool
- Pick a small task, like finding a library book.
- Ask who would use the tool and what they need.
- Think of one way the tool could make a mistake.
- Choose a test that could uncover that mistake.