Is Europe overregulating AI? | DW Business

Key Concepts

EU AI regulations, innovation, competitiveness, fragmentation of regulation, data protection, open-source AI, community notes, digital services act (DSA), freedom of speech, AI-created ads, metaverse, talent.

EU AI Regulation and Innovation

Julia Royce, Meta's public policy director for Central Europe, addresses concerns regarding the EU's new AI regulations. She notes that CEOs of major German tech companies like SAP and Siemens, along with the German government, have expressed reservations about the speed of implementation. Royce shares that the enforcement is supposed to be effective by early August and she believes it's simply too early. She argues that enforcing AI regulation prematurely would hinder innovation. She hopes that these critical voices are being heard. She suggests pausing and reassessing the effects of AI on industry and society in two years, rather than implementing regulations now.

Competitiveness and Fragmentation

Royce points to worries in Brussels about Europe's future competitiveness, referencing the Mario Draghi report from September of the previous year. The Draghi report encouraged innovation and new measures to regain competitiveness, emphasizing that small changes and doing "our homework" are needed, specifically fully harmonizing the digital market and addressing the fragmentation and overlap in existing regulations. She believes that many countries in Europe are suffering from recession right now also Germany and Brussels has woken up and there needs to happen something. She suggests that the omnibus procedure for simplifying rules could be a good opportunity for Brussels to step back and allow the industry to breathe.

Meta's Perspective and the Best-Case Scenario

From Meta's perspective, the best-case scenario is a two-year pause to reassess the AI landscape. The current uncertainty and regulatory fragmentation overwhelm companies, especially startups and mid-sized businesses, potentially discouraging them from developing AI applications due to legal risks. Meta can hire additional lawyers for their legal departments but how about startups? How about middlesiz companies? I mean they rather put the hands off them the application of AI applications which would be I think tragic because it's the future for many processes um rather than taking the legal risk and I think um there's a lot of things the the EU need to reconsider not to to hampering that innovation.

Data Protection and Legal Uncertainty

Royce highlights the issue of fragmented data protection regulations within the EU, citing the example of North Rhine-Westphalia's data protection agency attempting to block Meta's use of postings to train AI, despite Meta receiving a green light from the European Data Protection Board in December. She explains that Meta is on average eight months behind launching products in Europe because of legal uncertainty. She describes the difficulty in explaining to her US colleagues how one region can block something after European-wide approval, questioning the concept of a single European market. "Honestly, for training um our models with EU data or with data in within the EU, we have been waiting for over eight months. We are an average eight months behind. We're launching our products in Europe because of legal uncertainty."

Consent and Collaboration

Royce addresses criticisms that Meta is repeating mistakes from personalized advertising by not providing a clear consent option for AI training. She argues that there is a fundamental misunderstanding and emphasizes Meta's collaborative relationship with the EU Data Protection Board and the EU data protection authority in Dublin. She reiterates the need for member states to commit to decisions made at the EU level to avoid hindering the scaling of European startups.

Open-Source AI and Meta's Strategy

Royce confirms that Meta is committed to an open-source AI model, as advocated by Yann LeCun, Meta's head of AI research. While acknowledging competition from companies like OpenAI and Google's DeepMind, she states that Meta believes the open-source approach is the right way forward. She states that there is no debate about it. It is the open source model which is for us the way to to go for.

Policy Shift and the DSA

Royce addresses the perception of a radical shift in Meta's policy approach, particularly regarding fact-checking and community notes. She maintains that Meta adheres to the Digital Services Act (DSA) and that the switch to community notes was greenlit by the EU Commission. She also mentions streamlining content policy decisions. She states that she didn't get any push back and the more she explain it also to policymaker there was absolutely no negative um feedback by this.

Misinformation and Freedom of Speech

Royce asserts that Meta is a reliable partner in fighting misinformation and ensuring election safety, adhering to the DSA's requirements. She emphasizes that freedom of speech is a highly esteemed value in both the US and Europe, and Meta enables it within the guardrails of the DSA.

AI-Created Ads and Potential Concerns

Royce acknowledges the potential for AI-created ads to be a massive shift in the advertising market. While AI tools are already used by creators, she anticipates criticism regarding Meta being both the publisher and designer of ads, potentially leading to concerns about monopolies. She states that there will always be critics and people who are criticizing and a technology advancement.

The Metaverse and European Talent

Royce clarifies that the metaverse is still being developed, with significant activity in Germany, particularly in the industrial metaverse with companies like Siemens. She highlights Meta's presence in Europe through the acquisition of Oculus (now part of Meta Reality Labs) and emphasizes the importance of focusing on European talent in fundamental research and AI application. She concludes by advocating for allowing innovation to flourish without premature regulation.

Conclusion

The interview highlights the tension between fostering innovation in AI and regulating its potential risks within the EU. Meta advocates for a pause in AI regulation to allow for better understanding and to avoid hindering European competitiveness, particularly for startups. The discussion also touches on data protection, freedom of speech, and Meta's commitment to the DSA, while acknowledging potential challenges and criticisms related to AI-created ads and the company's overall policy approach.