EU AI Office publishes report on proposed criteria for independent AI model evaluators

The EU AI Office’s October 21 report covers proposed qualification and independence criteria for external evaluators of general-purpose AI models with systemic risk. It says experts from 19 organizations attended a July workshop and notes that its views may change. Separately, the AI Evaluator Forum says the office endorsed key AEF-1 provisions as a way for providers to meet independence provisions in the General-Purpose AI Code of Practice.