Introduction
The European Union is taking proactive steps to address systemic risks associated with general-purpose AI (GPAI) models. A workshop in Brussels, organized by the EU AI Office and the Network of Evaluators, aims to develop evaluation criteria in alignment with the EU AI Act and the General-Purpose AI Code of Practice.
Description
On April 28, a workshop will be held in Brussels to develop evaluation criteria for assessing systemic risk associated with general-purpose AI (GPAI) models. This initiative, organized by the EU AI Office and the Network of Evaluators, aims to enhance the scientific and practical evaluation of GPAI models, particularly those that may pose systemic risks, in line with the EU AI Act and the General-Purpose AI Code of Practice.
The Network of Evaluators, comprising research institutions, industry representatives, and government bodies, is focused on refining evaluation methodologies for GPAI to ensure compliance with regulatory frameworks. A previous workshop in December laid the groundwork for these evaluations and risk management strategies, informing the current mission of the Network.
The upcoming workshop will facilitate discussions on proportionality criteria for systemic risk evaluations, as defined in the AI Act and Code of Practice. Independent reviewers will assess recent evaluations against these criteria prior to the workshop, and participants will collaboratively explore insights and gaps in the evaluation process. A summary of the workshop will be produced, highlighting areas of alignment with the criteria and opportunities for refinement.
Organizations interested in contributing evaluations relevant to systemic risk under the AI Act are encouraged to submit their work through designated platforms. High-quality reviewers may also be invited to participate in the workshop.
Conclusion
The workshop represents a significant step in the EU’s efforts to ensure that GPAI models are evaluated rigorously and systematically. By aligning evaluation criteria with established regulatory frameworks, the initiative seeks to mitigate potential systemic risks, thereby fostering a safer and more reliable AI ecosystem. The outcomes of this workshop could have far-reaching implications for the development and deployment of AI technologies across various sectors.
References
https://digital-strategy.ec.europa.eu/en/events/eu-ai-office-hosts-person-workshop-towards-defining-best-practices-systemic-risk-evaluations
