Limited risk, meaning the system is permitted but must: Into which risk tier does such a system typically
A conversational chatbot with only disclosure duties sits in the limited-risk tier under the EU AI Act.
The question
Under the EU AI Act's risk-based classification, a customer-service chatbot that simply converses with users and poses no significant safety or rights risk is generally subject to transparency duties such as disclosing that the user is interacting with AI. Into which risk tier does such a system typically fall?
Preparing for AIGP? Take the free 5-min readiness quiz →
- Unacceptable risk, meaning the system is banned outright as a clear threat to core fundamental rightsPlausible but wrong: unacceptable-risk systems are prohibited; a benign chatbot is not banned.
- Limited risk, meaning the system is permitted but must meet transparency obligations toward users ✓Correct: a conversational chatbot falls in the limited-risk tier, where transparency duties such as AI-interaction disclosure apply.
- High risk, meaning the system must undergo conformity assessment and strict ongoing obligationsPlausible but wrong: high-risk status brings strict obligations and conformity assessment not triggered by a simple chatbot.
- Minimal risk, meaning the system carries no obligations at all under the risk frameworkPlausible but wrong: minimal risk carries no specific duties, but the disclosure duty places a chatbot in the limited-risk tier.
The trap
Confusing limited risk (transparency duties) with minimal risk (no obligations). How to remember it
A conversational chatbot with only disclosure duties sits in the limited-risk tier under the EU AI Act.
How many of these would you get right?
One of 499 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
- It is treated as a special category warranting heightened: Compared with ordinary personal data, how do data →
- Importer, meaning it places a non-EU provider's system: Under the EU AI Act's role framework, which role does →
- Intellectual property law: Which legal domain most directly governs whether that training use of the material →
- All 125 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.
