Skip to content
InfoRegulatory

AI Safety Newsletter #59: EU Publishes General-Purpose AI Code of Practice

Published
Record updated
View JSON

Summary

On July 10, 2025, the EU published the General-Purpose AI Code of Practice, a voluntary set of guidelines to help providers comply with the AI Act's GPAI obligations before they take effect on August 2, 2025. Its Safety and Security chapter applies only to providers whose models cross the Act's systemic-risk threshold of ≥10^25 FLOPs and requires them to identify, analyze and determine acceptable systemic risks throughout the model lifecycle.