Holding Silicon Valley Accountable Across the Atlantic
BRUSSELS — Four senior European Parliament deputies have drafted amendments to the EU’s 2024 AI Act that would make leading U. S. tech firms legally responsible for harm caused by their most advanced artificial-intelligence models. The move, revealed through internal correspondence, aims to close perceived loopholes in the bloc’s flagship AI governance framework.
Breaking news
Montenegro, Albania Seek NATO Help to Counter Russian Disinformation
Ronna McDaniel Weighs In on GOP's Path Forward
Market Sentiment Shifts as S&P 500 Snaps Three-Day Losing Streak
Voters Demand Stronger Congressional Oversight of Presidential PowerThe proposed changes would extend strict liability provisions to cover frontier AI systems developed by companies such as Google, Microsoft, and OpenAI. Under current rules, only providers classified as „high-risk” face detailed oversight, but the new language would impose broader obligations on developers of cutting-edge models regardless of their size or location.
MEPs argue that American corporations have benefited from European markets while avoiding the full weight of EU accountability standards. The amendments seek to ensure that any entity offering AI services to EU residents — particularly those deploying experimental or autonomous systems — must demonstrate compliance with rigorous safety and transparency benchmarks.
„This is about protecting European citizens from the unintended consequences of powerful AI,” said one of the drafting lawmakers, who requested anonymity while negotiations continue. „We cannot allow foreign companies to operate in regulatory gray zones.”
Can Global Tech Giants Comply With Stricter EU Rules?
The proposal also introduces mandatory impact assessments for frontier models and requires third-party audits before deployment. Companies failing to meet these standards could face fines of up to 6% of global annual turnover, mirroring penalties already embedded in the original AI Act.
Industry representatives warn that the expanded liability framework may stifle innovation and deter investment in European digital infrastructure. U. S. Trade Representative Katherine Tai has previously criticized similar measures as discriminatory toward American businesses.
However, European officials maintain that harmonized global standards are necessary to manage risks associated with rapid AI advancement. The European Commission is expected to review the amendments during upcoming trilogue talks between Parliament, Council, and executive bodies.
Frequently Asked Questions
If adopted, the revised regulations could reshape how multinational technology firms approach AI development and market entry within the EU. Legal experts suggest the outcome may influence parallel legislative efforts in the United States and other jurisdictions weighing similar oversight mechanisms.
What defines a „frontier” AI model under the proposal? The draft text refers to frontier models as highly capable systems exhibiting general-purpose functionality, including large language models and multimodal agents, which can perform tasks across diverse domains without task-specific training.
How would liability be enforced against non-EU companies? Enforcement would rely on existing GDPR-style mechanisms, including designated EU representatives, local legal proceedings, and reciprocal sanctions targeting assets or operations within European borders.
Will these changes delay AI innovation in Europe? Supporters argue that clear rules will boost long-term trust and adoption, while critics fear bureaucratic hurdles could slow deployment of next-generation AI applications by months or years.
