European Union: European Commission published final version of General-Purpose Artificial Intelligence Code of Practice, including testing requirement

On 10 July 2025, the European Commission published the final version of the General-Purpose Artificial Intelligence (GPAI) Code of Practice, including testing requirements. Under the safety and security chapter, providers of GPAI models with systemic risk must conduct testing and evaluation procedures aligned with Articles 55(1) and 56(5) of the AI Act, which outlines specific obligations for providers of general-purpose AI models with systemic risk. The Code highlights evaluations, including red-teaming, adversarial testing, and simulations, to assess systemic risks. It also requires evaluations to be reproducible and transparent, and inform risk acceptability decisions.