TRAIL #2: AI Ethics Made Simple

Vocabulary flashcards created from lecture notes on AI ethics, the FAIR test framework, common AI pitfalls, and responsible AI practices at PNB.

AI Ethics (according to International Business Machines)

Creating and using artificial intelligence in ways that help people - and avoid causing harm.

FAIR Test

Our pocket guide to ensure that AI is being used responsibly.

Fairness (FAIR Test)

Addresses the question: "Does the AI treat all users equally?"

Accountability (FAIR Test)

Addresses the question: "who is responsible when the AI makes mistake?"

Integrity (FAIR Test)

Addresses the question: "Is the AI system honest and reliable in its output?"

Respect (FAIR Test)

Addresses the question: "Does the system treat all individuals with dignity and avoid harmful outcomes?"

Bias

Happens when AI learns unfair patterns from the data it's trained on.

Deepfakes

Fake videos, images, or voices made by AI that look or sound real.

Black Box

When the system makes decisions, but no one really knows how or why.

Responsible Use at PNB

Simple steps for using AI responsibly at PNB: 1. Teach employees to spot problems, 2. Know who to tell, 3. Fix the problem quickly, 4. Learn and improve.