EU Tech Chief: US Will Adopt AI Guardrails Regardless
▼ Summary
– Henna Virkkunen argues that US and EU AI protections are converging despite different regulatory approaches.
– The EU relies on proactive legislation like the AI Act, while the US utilizes court cases and state laws.
– Michael Kratsios launched the Carolina Principles to discourage new AI regulations during the G20 summit.
– Recent enforcement actions against ChatGPT and Meta demonstrate practical similarities in platform restrictions.
– Regulatory asymmetries exist in speed and resources, but both systems are moving toward similar safety guardrails.
Henna Virkkunen, the European Union’s top tech official, asserts that the United States is inevitably moving toward comprehensive AI guardrails, regardless of current political rhetoric. Speaking to Axios at the G20 innovation ministerial in Chapel Hill, Virkkunen highlighted a growing convergence between European and American approaches to artificial intelligence regulation. While Europe relies on proactive legislative frameworks, the US often develops protections through litigation and state-level statutes. Despite these divergent paths, she argues that the practical outcomes are becoming increasingly similar.
“The differences are not so big when we look at it in practice,” Virkkunen stated. “In Europe, we have regulation in advance, and in the USA they are often coming with the court cases.”
This perspective stands in stark contrast to the stance taken by her hosts during the summit. Michael Kratsios introduced the Carolina Principles on September 1, urging G20 nations to avoid creating new AI-specific rules or regulatory bodies. Kratsios argued that policymakers “should not treat every emerging technology as a first-of-its-kind policy problem.” The event also featured high-profile industry leaders such as Elon Musk and Mark Zuckerberg, underscoring the intense scrutiny facing the tech sector.
Virkkunen contends that the US already possesses a robust regulatory landscape, albeit decentralized. She points to hundreds of existing rules at the state level that function as de facto regulations long before any federal court case arises. This claim is currently being tested by recent developments on both sides of the Atlantic.
On August 31, the European Commission officially designated ChatGPT, Reddit, and Roblox as very large online search engines under the Digital Services Act (DSA). These platforms were given four months to comply with the stringent requirements. Just one day later, the Commission issued its first information requests to over 30 AI providers under the AI Act. Commission spokesman Thomas Regnier noted that these questionnaires primarily focus on safety protocols and copyright issues.
In the United States, comparable regulatory pressure has emerged through the judicial system. In August, Meta reached a settlement with 51 state attorneys general, agreeing to implement strict usage limits for teenagers on Instagram and Facebook. These restrictions cap daily usage at two hours and enforce overnight blocks. The agreement carries a total value of $12.19 billion over ten years, covering defaults that European law expects platforms to proactively address.
Virkkunen views this dynamic as evidence of convergence, though she acknowledges the asymmetry in how rules are enforced. The EU’s AI Office utilizes a team of just 36 people to evaluate frontier models and send questionnaires to dozens of companies. Conversely, the US approach involves 51 attorneys general and settlements worth billions. Both systems generate binding rules, but they operate at different speeds and scales.
European momentum continues to build as well. Brussels agreed in May to adjust the implementation timeline of the AI Act, delaying high-risk obligations until December 2027. This strategic shift may contribute to Virkkunen’s assessment that the regulatory gap between the two regions is narrowing significantly.
(Source: The Next Web)
