Big Tech companies put pressure on the European Commission to water down the Code of Practice on General Purpose AI – which should help providers of AI models comply with the EU’s AI Act – a report by Corporate Europe Observatory (CEO) and LobbyControl published on Wednesday suggests.
Tech companies “enjoyed structural advantages” in the drafting process of the Code, and “weakened the rules around advanced AI", the groups found in a joint investigation based on internal interviews and analysis of lobby papers.
Thirteen experts appointed by the European Commission were tasked to draft the Code starting last September, using plenary sessions and workshops to allow some 1,000 participants to share feedback.
The CEO and LobbyControl research show that tech companies had more access to the drafting process than others.
"Model providers – companies developing the large AI models the Code is expected to regulate – were invited to dedicated workshops with the working group chairs. These workshops included 15 US based tech companies among which Google, Microsoft, Meta, Amazon, and OpenAI."
Other stakeholders, such as civil society organisations, publishers, and SMEs “had limited access, with participation during workshops mostly limited to emoji-based upvoting of questions and comments via the online platform SLIDO,” the report added.
The drafting process has been criticised, particularly by rights holders and publishers who fear the rules will contradict copyright law.
Delay
A spokesperson for the Commission confirmed earlier this week that it had received a letter from the US government’s Mission to the EU pushing back against the Code. The administration led by Republican President Donald Trump has been critical of the EU’s digital rules, claiming that it stifles innovation.
“The EU Commission’s obsession with ‘simplification’ and ‘competitiveness’ is opening the door to aggressive Big Tech lobbying. The Code of Practice is only among the first casualties of this single-minded focus on deregulation," said CEO researcher Bram Vranken.
Commission spokesperson Thomas Regnier told Euronews in a reaction that "No one received any ‘structural favouring’ throughout the process," and that all participants "had the same opportunity to engage in the process through the same channels."
The final version of the Code was set to come out early May but likely faces a delay.
The Commission did not comment on Monday whether or not the 2 May deadline would be met. However, both the guidelines on general-purpose AI and the final General-Purpose AI Code of Practice are expected to be published in May or June 2025, it said in a recently published consultation on GPAI guidelines.
In email to stakeholders, seen by Euronews, the Commission's AI Office said that the final text is expected to be published "ahead of August 2025", when the rules on GP AI tools enter into force.
The EU executive can decide to formalise the Code – which will be fully applicable in 2027 – through an implementing act.
This article has been updated to include a comment from the European Commission.
