Risk Assessment Tool for AI Act

This tool provides a potential risk-level for the use of AI systems. Select one or more concepts and then click the Check Risk Level button. You can test various combinations or try pre-selected examples below or select random values. The definitions of terms are provided at the document end.

Results

Definitions

Inputs are described in terms of the following concepts. For more details on the creation of the tool, methodology, and discussion - refer to the peer-reviewed publication.

Domain: the domain or sector or area within which the AI system is or will be deployed; e.g. Health, Education
Purpose: the purpose or end-goal for which the AI system is or will be used to achieve; e.g. Patient Diagnosis, Exam Assessment
Capability: the capability or application for what the AI system is or will be used to provide; e.g. Facial Recognition, Sentiment Analysis
User: the user or operator who is or will be using the AI system; e.g. Doctor, Teacher
Subject: the subject or individual or group which the AI system is or will be used towards; e.g. Patients, Students

The outputs represent a potential risk level that represents whether the selected combination satisfies the AI Act Annex III:

High-risk means there is an exact match
Medium-risk means the combination can lead to a potential match (e.g. when the Purpose can take Any value)
Low-risk means the output does not match any clause (i.e. not high-risk) and does not have the potential to match any cause (i.e. not medium-risk).

Acknowledgements

This implementation is by Harshvardhan J. Pandit based on the below work. Source code is available via GitHub repo under a permissive license.

Cite this work as: Delaram Golpayegani, Harshvardhan J. Pandit, and Dave Lewis. "To Be High-Risk, or Not To Be—Semantic Specifications and Implications of the AI Act’s High-Risk AI Applications and Harmonised Standards" Proceedings of the 2023 ACM Conference on Fairness, Accountability, and Transparency. 2023. https://doi.org/10.1145/3593013.3594050