Ethics and bias in AI in Intro to Computing - Real World Applications

Real World Mode - Ethics and bias in AI

Ethics and Bias in AI: The Librarian Analogy

Imagine a large public library where a librarian helps visitors find books and information. This librarian represents an AI system. The librarian tries to be helpful and fair, but sometimes the books on the shelves reflect the opinions and history of the people who wrote them. If the librarian only recommends certain books or ignores others, visitors might get a one-sided view. This is like bias in AI. Ethics in AI is like the library's rules to make sure the librarian treats everyone fairly, respects privacy, and provides balanced information.

Mapping AI Ethics and Bias to the Librarian Analogy

Computing Concept	Real-World Equivalent
AI System	The librarian helping visitors find information
Training Data	The collection of books and materials the librarian uses
Bias in AI	The librarian favoring certain books or topics due to the collection's limitations or past choices
Ethical Guidelines	The library rules ensuring fair treatment, privacy, and balanced recommendations
Algorithm Decisions	The librarian's choices on which books to suggest or prioritize
Feedback and Correction	Visitors pointing out missing or unfair recommendations, leading to library updates

A Day in the Library: Understanding AI Ethics and Bias

One day, a visitor asks the librarian for books about local history. The librarian quickly suggests several books but mostly from one perspective because the library's collection is richer in those. Another visitor notices this and tells the librarian about missing viewpoints. The librarian then works with the library team to add more diverse books and update the recommendation approach. This process shows how AI systems can unintentionally be biased and how ethical practices help improve fairness and accuracy over time.

Where the Librarian Analogy Breaks Down

• The librarian is a human with feelings and intentions, while AI is a machine following programmed rules without consciousness.
• The library collection is physical and updated manually, whereas AI training data can be vast and updated through complex automated processes.
• Visitors can directly talk to the librarian, but users may not always understand or influence AI decisions easily.
• The analogy simplifies the technical complexity of algorithms and data processing involved in AI.

Self-Check Question

In our analogy, what would the library rules be equivalent to in AI?

Key Result

Ethics and bias in AI are like a librarian following fair library rules to recommend balanced books without favoritism.