EU's rules for AI developers take effect
Big Tech is complaining
The AI Act will finally come into effect in the EU this week. A voluntary code of practice designed to inform GenAI developers of how best to stay within the law was only published on 10th July. Meta has refused to sign it, and others are doing so reluctantly.
The European Union's long-anticipated AI Act will take effect this week, as new rules targeting developers of general-purpose artificial intelligence (GenAI) systems officially come into force on 2nd August.
While regulators hail the legislation as a global benchmark for safe and ethical AI deployment, major tech firms are pushing back – accusing Brussels of overreach, poor timing, and potentially stifling innovation.
The AI Act, first proposed in 2021 and enacted in March 2024, is designed to regulate a fast-evolving technology landscape, one that exploded with the launch of OpenAI's ChatGPT in late 2022.
Since then, generative AI has reshaped enterprise software, marketing, and creative industries, sparking both unprecedented innovation and legal challenges, particularly around copyright infringement.
The new rules apply to developers of GenAI systems such as ChatGPT and Meta's Llama, and require them to evaluate models, mitigate systemic risks, report major incidents, conduct adversarial testing, maintain cybersecurity, and even disclose the energy efficiency of their systems.
These obligations, which fall under the AI Act's tiered rollout, are only the beginning. More stringent rules for so-called "high-risk" AI applications – those affecting health, safety, rights, and democracy – will follow in August 2026.
To ease the compliance burden, the European Commission recently introduced a voluntary Code of Practice, drawn up by 13 independent experts. This framework, though not legally binding, is meant to serve as a blueprint for how GenAI developers can align with the Act's obligations.
It includes requirements such as issuing summaries of training data, ensuring alignment with EU copyright law, and engaging with the new EU AI Office overseeing enforcement.
However, the Code rollout has been less than smooth. Originally due in May, the guidance was only published on 10th July, leaving developers less than a month to absorb and implement it before compliance expectations kick in.
That short runway has drawn sharp criticism from major tech companies, most vocally Meta.
The social media giant, developer of Llama large language model (LLM), has refused to sign the Code, with Joel Kaplan, Meta's chief global affairs officer, declaring on LinkedIn that "Europe is heading down the wrong path on AI."
Meta argues that the Code introduces legal uncertainty beyond what is stipulated in the Act.
Others, including legal experts, have raised similar concerns.
Nils Rauer, partner at law firm Pinsent Masons, told The Register: "They've tried their best, but it remains quite generic on a number of issues, including copyright issues. You need to be generic to some extent, but you need to cover a lot of cases, a lot of situations, in order to be able to make that guidance precise and practical. It's a difficult task but we had hoped for more specific, more practical guidance."
Monika Sobiecki, partner at law firm Bindmans, raised concerns about timing and procedural completeness.
"The guidance hasn't fully gone through the trialogue [between the European Parliament, Council, and Commission]," she said.
"Although we've got most of the shape of it, you would have expected it to be approved by the European Commission and parliament before it is implemented, and compliance is expected."
Nevertheless, some Big Tech players are opting in, albeit with caveats.
Google, via a blog post by Kent Walker, its global affairs president, confirmed it will sign the Code. But Walker warned that the AI Act, in its current form, risks impeding Europe's AI competitiveness.
"Departures from EU copyright law, steps that slow approvals, or requirements that expose trade secrets could chill European model development," he said.
Microsoft has also signalled its intention to sign, while smaller GenAI startups and European developers are scrambling to understand the full implications of the new requirements.