In red teaming, experts deliberately try to make an AI fail or give harmful answers, to find weaknesses before launch.
Example
A team tries to get a customer service bot to reveal internal discount codes.
For pros
Tests cover jailbreaks, prompt injection, data leaks and dangerous content, among others. The EU AI Act requires adversarial testing for GPAI models with systemic risk. Findings feed into guardrails, training and evals.
