Governance, Law & Compliance
Risk tier
A category used to group AI systems by the level of potential harm or regulatory concern.
Example
A regulation assigns stronger obligations to an AI system placed in a higher risk tier.
Why it matters
It helps apply more demanding checks where potential consequences justify them.
You'll hear it
“High risk under which framework, and for which use?”
Can you control it?
Sometimes
Sometimes. An organization can define or influence a risk tier through its policies, processes, contracts, and product choices. Individual users may have fewer controls.
Common questions
- Can the same model fall into different tiers?
- Yes. Classification can depend on the application and context rather than the model alone.
- Does a low tier mean no risk?
- No. A broad classification does not rule out problems in a particular situation.
- Can the tier change later?
- Yes. A changed use or revised assessment may alter the classification under the applicable framework.