We are pleased to announce that Technopolis Group has been selected by the European Commission’s AI Office to work on sociotechnical risk assessment for general-purpose AI models. Alongside collaborators we will be developing risk models and tailored evaluations to support the enforcement of the AI Act. The project builds on Technopolis Group’s prior work in supporting the AI Office with the preparation of the first Code of Practice for General-Purpose AI (GPAI), as part of the implementation of the AI Act.

This work, which is technical, legal and policy relevant, will shape how AI systems are evaluated on sociotechnical risks. Specifically, we will develop a risk-modelling approach; provide tools to identify and assess potential risks associated with general-purpose AI; design a technical compliance toolkit; establish a risk monitoring framework with regular reporting; and carry out ad hoc evaluations as required. 

Technopolis Group is excited to be working alongside a strong group of collaborators like TNO, Algorithm Audit, Wik-Consult, and Spark, as well as academic experts from Leiden University and the University of Oxford. 

The study formally commenced with an inception meeting in January, with project implementation running through to the end of 2028.