Reliability
Domain: Safety & Ethics
🎯 Motivation
The AI consistently produces correct, safe, and expected results. Unreliable AI is dangerous.
📋 Concrete Example
Reliability: ‘In 1000 tests, the AI gave correct answers 98% of the time, never gave dangerous advice, and always handled errors gracefully.’ High reliability.
🔗 Analogy
A reliable car - it starts every morning, does not break down, and gets you to school safely. Reliability in AI means consistent, dependable performance.