Definition
A public document published by an AI developer that describes a model's capabilities, known risks, safety evaluations, and what safeguards are in place before it is released. Think of it as the safety data sheet for an AI model — a structured disclosure of what the system can do and what could go wrong.
Why it matters
System cards are becoming a de facto transparency standard, increasingly referenced by regulators and enterprise buyers when assessing AI risk. A company deploying a third-party AI model should review its system card the same way a manufacturer reviews a component's safety certification.