AI Safety & Responsible AI
System card
A broader report describing the behavior, testing, and risks of an entire AI system.
Example
A system card explains how the complete product was tested, not only its underlying model.
Why it matters
It helps readers assess risks created by the complete product, including its connected parts.
You'll hear it
“This report covers the assistant with its tools enabled.”
Can you control it?
Sometimes
Sometimes. Some products or workflows let you influence a system card, while others keep it fixed or manage it automatically. Check the settings and documentation for the tool you are using.
Common questions
- Is a system card the same as a model card?
- No. The wider product can behave differently because of its tools, instructions and safeguards.
- Can a published report leave out important details?
- Yes. Its scope and level of disclosure may not cover every concern.
- Does a good test result rule out later incidents?
- No. New users, changing services or untested situations can still expose problems.