EU Tech Chief: AI Guardrails Are Coming to the US Anyway
Henna Virkkunen, the EU’s tech chief, argues that despite the US administration’s stance on AI regulations, guardrails are inevitable. She suggests that both the EU and US ultimately reach similar AI protections through different means: the EU regulates in advance, while the US relies on court cases and state laws.
Virkkunen’s Points:
- The differences between EU and US AI regulations are minimal in practice.
- The US already has AI rules at the state level, even without federal regulations.
- She cites examples of recent actions: the EU Commission designating ChatGPT as a large online search engine and sending AI providers questionnaires under the AI Act, while Meta settled with 51 US attorneys general over teen usage restrictions.
Price of Convergence:
- The Meta settlement is valued at $12.19 billion over ten years.
- The EU’s AI Office has a team of 36 evaluating frontier AI models, while the US deals with 51 attorneys general and a massive settlement.
Bridging the Gap:
- Brussels agreed in May to delay high-risk obligations under the AI Act to December 2027, indicating movement from the European side.
- Virkkunen acknowledges the gap is small, attributing it to both sides working towards similar AI protections.