#AI horizons 25-02 – Legal and compliance and news

As artificial intelligence continues to reshape industries across the global economy, regulatory frameworks are rapidly evolving to address both the opportunities and challenges this technology presents. The emerging legal landscape for AI reflects competing priorities between innovation and oversight, with significant implications for businesses deploying these technologies. Recent developments in the European Union, international summits, and stakeholder reactions indicate a complex and sometimes fractured approach to AI governance that business leaders must navigate.

EU AI Act Implementation Timeline and Requirements

The European Union’s AI Act has now entered its enforcement phase, establishing the world’s most comprehensive regulatory framework for artificial intelligence. Beginning February 2, 2025, the Act immediately prohibits several AI applications deemed high-risk:

• Social rating systems
• Predictive policing AI for individual profiling
• Emotion recognition technology in workplaces and schools
• AI systems exploiting vulnerabilities or using manipulation techniques
• Real-time facial recognition in public spaces (with limited law enforcement exceptions)
• Biometric categorization systems for identifying personal characteristics

The implementation schedule follows a graduated approach:

• From August 1, 2025: General-purpose AI model providers must provide transparency regarding technical documentation and training data
• Major AI models will require security audits
• Subsequent phases will extend regulations to high-risk AI applications across infrastructure, education, employment, banking, justice, and other critical sectors

Enforcement will be managed by the new AI Office along with national authorities. Violations carry substantial financial penalties:

• Up to 7% of global sales for prohibited practices
• Up to 3% of global sales for other infringements

To support compliance, the European Commission has published non-binding guidelines explaining the definition of AI systems under the Act (https://digital-strategy.ec.europa.eu/en/library/commission-publishes-guidelines-ai-system-definition-facilitate-first-ai-acts-rules-application). These guidelines aim to help providers determine whether their software qualifies as AI under the new legislation and will evolve based on practical experience and emerging use cases.

Global Divergence on AI Regulation

The recent AI Action Summit in Paris highlighted significant international divergence on AI governance approaches, marking a shift from previous summits that focused primarily on AI safety concerns:

• The summit emphasized reducing regulatory burdens to foster innovation
• The United States advocated for a pro-business stance, with Vice President JD Vance criticizing Europe’s “excessive regulation” of AI
• Both the UK and US declined to sign the Paris AI summit declaration promoting ethical, transparent, and sustainable AI development
• Only 26 of 60 countries agreed to restrictions on military AI applications
• China and several European nations endorsed collaborative AI development approaches

This fractured global approach creates a challenging environment for multinational businesses, which may face different compliance requirements across jurisdictions. The US position in particular signals resistance to international standards that might constrain American technological dominance.

Investment vs. Regulation: The Competitive Landscape

A notable pivot occurred at the Paris summit, with European stakeholders shifting focus from strict regulatory measures toward investment strategies to compete globally:

• France pledged approximately $114 billion toward AI research, startups, and infrastructure
• The EU announced a roughly $210 billion initiative to strengthen Europe’s AI capabilities and technological self-sufficiency
• France allocated 1 gigawatt of nuclear power specifically for AI development
• The European Commission withdrew a proposed “liability directive” that would have made it easier to sue companies for AI-related harms

This recalibration reflects growing concern that excessive regulation could hamper Europe’s competitiveness against less-regulated markets like the US and China. Business leaders, including Capgemini CEO Aiman Ezzat, have publicly criticized the EU’s regulatory approach as having gone “too far,” potentially hindering global companies’ ability to deploy AI technologies within the region.

Industry Pushback on Voluntary Frameworks

Beyond formal regulations, voluntary frameworks and codes of practice are also facing significant industry resistance:

• Google’s senior public affairs official Kent Walker described the EU’s proposed voluntary code of practice for advanced AI models as a “step in the wrong direction” for European competitiveness
• Meta’s chief lobbyist Joel Kaplan called the code’s requirements “unworkable and technically unfeasible” and indicated Meta would not sign in its current form
• The voluntary code would apply to companies operating general-purpose AI models including OpenAI, Anthropic, Google, Meta, and Microsoft
• Contentious areas include training data disclosure, management of systemic risks, copyright considerations, and third-party model testing

This resistance highlights tensions between regulatory objectives and practical implementation challenges, particularly for complex AI systems.

Supporting Compliance: AI Literacy Initiatives

To facilitate implementation of regulations, the European Commission has established resources to enhance AI literacy among stakeholders (https://digital-strategy.ec.europa.eu/en/library/living-repository-foster-learning-and-exchange-ai-literacy):

• Article 4 of the AI Act, effective February 2, 2025, requires AI providers and deployers to ensure sufficient AI literacy among staff and users
• The EU AI Office has compiled a “living repository” of AI literacy practices from organizations participating in the AI Pact
• The repository showcases various practices categorized by implementation status: fully implemented, partially rolled out, or planned
• While replicating these practices doesn’t guarantee compliance, the repository serves as a resource to foster learning and exchange among AI stakeholders

Ethical Dimensions and Institutional Perspectives

Beyond government regulation, institutional stakeholders are contributing to the AI governance discourse:

• The Vatican has released “ANTIQUA ET NOVA: Note on the Relationship Between Artificial Intelligence and Human Intelligence”, (https://www.vatican.va/roman_curia/congregations/cfaith/documents/rc_ddf_doc_20250128_antiqua-et-nova_en.html) addressing ethical implications of AI development
• This represents continued engagement from the Catholic Church, which has been working on AI ethics since at least 2016
• Such institutional perspectives may influence public opinion and regulatory approaches, particularly regarding ethical boundaries and human-centric AI development

Why It Matters: Strategic Implications for Business Leaders

The evolving legal landscape for AI presents both challenges and opportunities that demand strategic attention from business leaders:

1. Regulatory Arbitrage Considerations: Varying regulatory regimes create potential advantages for strategic operational placement, but also compliance complexity for global operations
2. Compliance Investment Planning: Organizations must budget for significant compliance resources, particularly for operations in the EU, with potential penalties reaching up to 7% of global sales
3. Competitive Innovation Balance: Finding the optimal balance between regulatory compliance and competitive innovation will be critical, particularly as regions compete through differing approaches
4. Ethical Leadership Opportunity: Forward-thinking organizations can differentiate through ethical AI deployment that anticipates regulatory trends rather than merely reacting to them
5. Stakeholder Engagement Strategy: Actively participating in the development of voluntary codes and frameworks may help shape more practical and business-friendly approaches

Business leaders should establish cross-functional AI governance teams that include legal, technical, ethical, and business strategy expertise to navigate this complex landscape effectively. Regular monitoring of regulatory developments and adjustment of AI development roadmaps will be essential for maintaining both compliance and competitive advantage in the rapidly evolving AI ecosystem.