Brussels is pressing ahead with the enforcement of its landmark AI Act, which includes new guidance on prohibited AI practices, despite threats from former US President Donald Trump regarding the regulation of American tech companies.
The AI Act is seen as the world’s most comprehensive AI regulation. On Sunday, the European Union began enforcing provisions that ban certain practices, including the creation of facial recognition databases through internet scraping. New guidance on how these rules should be applied will be released by the European Commission on Tuesday, with further provisions targeting high-risk AI applications, such as those in healthcare, to be rolled out by 2027.
Follow THE FUTURE on LinkedIn, Facebook, Instagram, X and Telegram
The EU’s push for enforcement comes as US-based tech companies, supported by Trump’s administration, express concerns over the regulation. Trump has warned that the EU’s treatment of American firms could result in retaliation, particularly regarding fines imposed on companies like Meta and Google. Trump’s administration has also signaled a shift in the US stance on AI regulation, promoting a less restrictive approach, including the announcement of a $500 billion AI infrastructure project backed by SoftBank and OpenAI.
Despite this pushback, the European Commission is steadfast in its commitment to enforcing the AI Act. The law requires companies developing high-risk AI systems to be more transparent about their processes and undergo risk assessments. Non-compliance could lead to hefty fines or even a ban from the EU market.
Big Tech has raised concerns that the EU’s transparency requirements could stifle innovation, particularly rules allowing third-party inspections of AI models for risk assessments. Meta has been vocal about the “onerous” nature of these provisions. However, Brussels continues to assert its position as the global leader in trustworthy AI, even as it navigates increasing opposition from the tech sector.
Caterina Rodelli, an EU policy analyst at Access Now, suggests that the approach to implementing the AI Act could shift under the new US administration. “There’s a risk that regulators could relax the rules, potentially undermining their effectiveness,” she noted.
While the EU’s recent bans have been clear, much is still to be determined in negotiations over the Code of Practice for general-purpose AI, which will affect major AI models such as Google’s Gemini and OpenAI’s GPT-4. These discussions, which involve hundreds of stakeholders, are set to conclude in April.