News

AI Regulation Takes Shape: Global Policy Landscape in December 2023

December 22, 2023 3 min read

As 2023 closes, the regulatory landscape for AI has transformed dramatically. From the EU’s comprehensive legislation to the US executive actions and global cooperation efforts, governments worldwide are actively shaping how AI will be developed and deployed.

EU AI Act: The Global Benchmark

The European Union’s AI Act represents the world’s most comprehensive AI legislation, with final negotiations completed in December 2023.

Risk-Based Framework

The Act categorizes AI systems by risk level:

Unacceptable Risk (Banned)

  • Social scoring by governments
  • Real-time biometric surveillance in public spaces
  • Manipulation of vulnerable groups
  • Predictive policing based on profiling

High Risk (Heavy Regulation)

  • Employment and worker management
  • Education and vocational training
  • Critical infrastructure
  • Law enforcement applications
  • Immigration and border control

Limited Risk (Transparency Required)

  • Chatbots must disclose AI nature
  • Emotion recognition systems
  • Deepfake content labeling

Minimal Risk (Light Touch)

  • AI-enabled video games
  • Spam filters
  • Most business applications

Foundation Model Requirements

Special provisions for general-purpose AI:

  • Technical documentation requirements
  • Transparency about training data
  • Compliance with copyright law
  • Safety testing and reporting

Penalties

  • Up to 35 million euros or 7% of global revenue
  • Graduated based on severity
  • Enforcement by national authorities

US Executive Order on AI

President Biden’s October 2023 executive order established significant new requirements:

Safety and Security

  • Red-team testing requirements for powerful models
  • Reporting to government before deployment
  • Standards development through NIST
  • Watermarking and content authentication

Civil Rights Protection

  • Guidelines against algorithmic discrimination
  • Protections for workers
  • Privacy safeguards
  • Transparency requirements

Innovation Support

  • National AI Research Resource pilot
  • Immigration reforms for AI talent
  • Small business support programs
  • Research funding increases

Implementation Timeline

  • 90-day requirements for initial reporting
  • 180 days for standards development
  • 270 days for agency guidance
  • Ongoing monitoring and updates

Global Regulatory Developments

United Kingdom

  • Pro-innovation approach emphasizing flexibility
  • Sector-specific regulation rather than horizontal
  • AI Safety Institute for frontier model testing
  • Voluntary commitments from companies

China

  • Algorithmic recommendation regulations
  • Deep synthesis (deepfake) rules
  • Generative AI provisional measures
  • Focus on content control and social stability

Canada

  • Artificial Intelligence and Data Act (AIDA) proposed
  • Focus on high-impact AI systems
  • Ministerial authority for standards
  • Currently in legislative process

Japan

  • Light-touch regulatory approach
  • Industry self-governance emphasized
  • Focus on promoting innovation
  • Soft law guidance preferred

Industry Response

Compliance Preparations

Companies are actively preparing:

  • Establishing AI governance teams
  • Conducting system inventories
  • Developing compliance frameworks
  • Training staff on requirements

Industry Associations

Groups advocating for workable regulations:

  • BSA (The Software Alliance)
  • TechNet
  • Information Technology Industry Council
  • Partnership on AI

Self-Regulatory Initiatives

Voluntary commitments including:

  • Frontier Model Forum
  • Content authenticity initiatives
  • Safety testing frameworks
  • Transparency reporting

Key Debates Ongoing

Open Source Treatment

  • Should open-source models have different rules?
  • How to balance innovation with safety?
  • Who bears responsibility for model misuse?

Definition Challenges

  • What constitutes an “AI system”?
  • How to measure model capability?
  • When does a model become “high risk”?

Enforcement Practicality

  • How to audit complex AI systems?
  • Cross-border enforcement challenges
  • Resource requirements for regulators

Looking Forward

2024 regulatory priorities:

  • EU AI Act implementation details
  • US agency guidance releases
  • International coordination mechanisms
  • Technical standards development

The regulatory landscape continues evolving rapidly, requiring organizations to stay informed and adaptable as requirements take shape.