AI Testing

Definitions

1. The broader practice of validating AI system behavior through test sets, scenario runs, regression suites, manual review, and production monitoring. AI testing differs from ordinary software testing because outputs can be probabilistic and quality-dependent rather than strictly deterministic.

   In plain English: The process of checking whether an AI system behaves correctly and usefully.

   Example: "Their AI testing pipeline mixed automated scoring with weekly human review to catch subtle tone and policy regressions."

Related Terms

• Eval Suite
• AI Metric
• AI Performance
• AI Quality
• AI Simulation
• AI Test
• Human Baseline
• LLM Benchmark
• Model Eval
• Prompt Evaluation
• Prompt Testing
• Safety Evaluation
• Safety Testing
• Truthfulness