A long-awaited set of voluntary rules for general-purpose artificial intelligence (GPAI) has been delayed, but the European Commission says it will be published before August.
US tech companies including Amazon, IBM, Google, Meta, Microsoft and OpenAI are urging the EU to keep the Code of Practice “as simple as possible”. They raised concerns during a meeting with EU official Werner Stengg, saying the rules should avoid “redundant reporting” and limit extra administrative work.
The Code aims to guide AI model providers, such as makers of tools like ChatGPT, in complying with the EU’s AI Act. Originally expected on 2 May, the final draft was postponed after requests for more time to give feedback.
Thirteen experts have been working on the guidelines since last September. Around 1,000 participants have joined workshops and meetings to offer input. However, earlier versions of the Code drew criticism. Publishers warned of risks to copyright, while tech giants feared it could hinder innovation.
The companies stressed the final version should not go beyond the AI Act and should allow time for proper implementation.
Meanwhile, concerns from the creative sector continue. ABBA’s Björn Ulvaeus recently warned EU lawmakers that proposals pushed by Big Tech could weaken artists’ rights. He leads CISAC, which defends global author and composer rights.
The Commission confirmed that the draft Code should be released before summer, ahead of 2 August when GPAI rules begin applying. The full AI Act will roll out by 2027.
