In October 2023, the White House mandated safety standards for AI systems. Yet, a recent Deloitte Global AI Survey 2023 found 68% of tech executives believe current governance frameworks are insufficient to manage emerging risks.
Authorities focus on managing risks and establishing sound practices for AI adoption. However, the technology's rapid evolution and complex societal impacts outpace regulatory responses. The Brookings Institute reports that AI model development, with new capabilities emerging monthly, outstrips legislative cycles of several years.
Given these fragmented efforts and accelerating AI development, a truly comprehensive, globally harmonized AI governance framework appears elusive. This will likely lead to continued reactive policymaking. The global scramble for national AI regulations creates an ineffective, fragmented oversight landscape, encouraging regulatory arbitrage and making comprehensive safety and ethical governance impossible.
The Patchwork of Global AI Governance
- The EU's AI Act, expected in 2024, categorizes AI systems by risk, imposing strict rules on 'high-risk' applications (European Commission).
- The UK invested £100 million in an AI Safety Institute to research frontier models (UK Dept. for Science, Innovation & Technology).
- The US National Institute of Standards and Technology (NIST) released its voluntary AI Risk Management Framework in January 2023 (NIST).
- Leading AI developers like OpenAI and Google DeepMind committed to external audits of their powerful models (AI Frontier Model Forum).
These varied efforts from governments and industry leaders establish foundational frameworks, yet the absence of unified global standards creates a fragmented oversight landscape. This piecemeal approach, while acknowledging AI's profound implications, risks creating loopholes and inconsistent protections.
