EU AI guardrails fall short on ambition, execution, study warns
This digest was compiled by AI from multiple sources — links to the originals are below.
A study in Big Data & Society warns that the EU's AI 'guardrails' lack ambition and execution, and are too rigid to keep pace with rapid technological change. The findings challenge the bloc's regulatory approach as AI development accelerates globally.
Regulatory Gaps
The study, published July 16, argues that the EU's AI Act and related measures fail to address key risks, including bias, surveillance, and environmental impact. Researchers from the University of Cambridge and other institutions found that the rules are too slow to adapt to new AI capabilities, such as generative models. The authors call for a more dynamic framework that can evolve with the technology.
Implementation Challenges
The EU's approach relies on ex-ante conformity assessments and human oversight, but the study says these mechanisms are already outdated. For example, the requirement for human review of AI decisions may be impractical for high-speed trading systems or autonomous vehicles. The paper notes that only 37% of EU companies have adopted AI, suggesting the rules may also stifle innovation without providing adequate protection.
What's Next
The European Commission is set to review the AI Act in 2028, but critics argue that the pace of change demands earlier revisions. It remains unclear whether the bloc can balance innovation with its stated goal of 'trustworthy AI'.
1 source
EU AI guardrails fall short on ambition, execution, study warns
