Performing state-of-the-art model evaluation: Beyond the baseline duties owed by all GPAI providers, which
Systemic-risk GPAI providers additionally must run state-of-the-art evaluations including adversarial testing and mitigate systemic risks.
The question
A provider's general-purpose AI model has been designated as posing systemic risk. Beyond the baseline duties owed by all GPAI providers, which additional obligation applies specifically because of the systemic-risk designation?
Preparing for AIGP? Take the free 5-min readiness quiz →
- Maintaining technical documentation and providing information to downstream providers that integrate the modelPlausible, but documentation and downstream information are baseline GPAI duties owed by all providers, not extra ones.
- Putting in place a policy to comply with applicable Union law on copyright and related rights throughout model trainingA copyright policy is a baseline GPAI obligation for every provider, not an add-on for systemic-risk models.
- Performing state-of-the-art model evaluation, including adversarial testing to identify and mitigate systemic risks ✓Correct: adversarial testing and systemic-risk mitigation are additional duties tied to the systemic-risk designation.
- Publishing a sufficiently detailed summary that describes the content used to train the model in questionThe training-content summary is also a baseline GPAI obligation, so it is not specific to systemic-risk models.
The trap
Selecting a baseline GPAI duty (documentation, copyright policy, training summary) as though it were a systemic-risk-only obligation. How to remember it
Systemic-risk GPAI providers additionally must run state-of-the-art evaluations including adversarial testing and mitigate systemic risks.
How many of these would you get right?
One of 1581 AIGP questions on Certsqill. Take a free five-minute check and see your score per domain — not one number, but which section to open tonight.
Test your AIGP readiness — freeMore Understanding How Laws, Standards and Frameworks Apply to AI questions
- Give clinicians the tool's confidence and key reasoning so: Which design measure best achieves meaningful →
- The first faces the top fine tier for a prohibited: Under the EU AI Act's fine structure, how do their maximum →
- Deployer, because it uses an AI system under its own: Under the EU AI Act, which operator role does the →
- All 394 Understanding How Laws, Standards and Frameworks Apply to AI questions →
Part of the Certsqill AIGP question bank · Understanding How Laws, Standards and Frameworks Apply to AI ·
Every answer, right and wrong, comes with its own explanation.