The world is splitting into two camps on AI governance, and the stakes are higher than ever. While the Trump administration argues that heavy regulation will stifle innovation, regulators in the European Union, the United Kingdom, and the US health sector are quietly building comprehensive frameworks to govern artificial intelligence systems. This divergence is playing out in real time at international forums, in legislative chambers, and in the practical decisions that companies must make about how to deploy AI safely.

Why Is the US Pushing Back Against AI Regulation?

At the G20 Innovation Ministerial held in Chapel Hill, North Carolina, in early September 2026, the US government made its position crystal clear: regulation should not slow down AI development. Michael Kratsios, director of the Trump administration's Office of Science and Technology Policy, warned that "slow policy adaptation makes it harder to capture the full benefits of innovation" and that "regulations built for a different technological era prove poor fits for rapidly evolving, generally applicable inventions".

The Trump administration is promoting what it calls the "Carolina Principles," a framework designed to prioritize industry-led standards and commercialization pathways over government oversight. David Sacks, the Trump administration's former AI czar, went further, calling it a "disaster" if the US adopted an "FDA for AI." He argued that any regulatory agency approving AI releases would become "like a DMV for AI, where new models get stuck in a long queue waiting for approval." Sacks contended that AI is "just too dynamic and fast-moving for that type of approach".

This position puts the US government at odds with Anthropic CEO Dario Amodei, who has advocated for exactly that kind of government review process. Amodei has written that "frontier AI models, like airplanes, should be required to go through technical testing and auditing, and their release should be blocked or reversed as a threat to public safety if they do not meet high standards of safety".

What Are Europe and Health Regulators Actually Building?

Meanwhile, on the other side of the Atlantic, regulators are moving in the opposite direction. The European Union's Digital Omnibus introduced the first substantive amendments to the AI Act in July 2026, with policymakers focusing on how AI requirements will interact with existing regulatory frameworks. The UK's Medicines and Healthcare products Regulatory Agency (MHRA) published guidance documents addressing when AI-enabled products qualify as medical devices and how they should be regulated, including new guidance on ambient voice technologies and digital mental health tools.

In the US health sector, the momentum for AI governance is accelerating. The Senate Health, Education, Labor and Pensions (HELP) Committee unanimously approved the Health Information Privacy Reform Act on July 30, 2026, which would impose privacy and security standards comparable to HIPAA (Health Insurance Portability and Accountability Act) on non-HIPAA-regulated digital health companies that collect "wellness data" like step counts and vital statistics.

Senator Mark Warner released a comprehensive AI policy framework in July 2026 that includes several legislative proposals. Most notably, the AI Artificial Intelligence Access, Gatekeeper Exchange, and Nondiscriminatory Transfer (AGENT) Act would establish privacy, cybersecurity, and interoperability requirements for consumer-facing AI agents, while the Secure Artificial Intelligence Development Act would require pre-deployment testing and create a voluntary safety incident reporting system for advanced AI models.

How Are Regulators Addressing AI in Healthcare Specifically?

The FDA's Digital Health Center of Excellence opened public comment on a discussion paper proposing a risk-based, "competency-based" framework for evaluating generative AI-enabled medical devices across their premarket and post-market lifecycle. This represents a practical approach to governing AI in one of the most sensitive sectors: healthcare.

The urgency behind these moves became clear when the Federal Trade Commission (FTC) sued telehealth provider Hims & Hers Health on July 29, 2026, alleging that the company misled consumers about subscription practices and undisclosed sharing of health data with third-party advertising platforms through tracking tools like the Meta Pixel. The same month, a $36.4 million False Claims Act settlement targeted unnecessary genetic testing kickback schemes, underscoring continued scrutiny of both privacy practices and billing integrity in digital health.

Steps Regulators Are Taking to Govern AI Across Sectors

• Privacy Protections: Extending HIPAA-like standards to non-HIPAA-covered digital health companies that collect wellness data, requiring advance notice and opt-out opportunities for consumers.
• Pre-Deployment Testing: Requiring advanced AI models to undergo safety testing before release and establishing voluntary incident reporting systems for developers.
• Medical Device Classification: Clarifying when AI-enabled products qualify as medical devices and establishing risk-based evaluation frameworks for their approval and monitoring.
• Consumer-Facing AI Requirements: Setting privacy, cybersecurity, and interoperability standards for AI agents that interact directly with consumers.
• Cross-Regulatory Coordination: Ensuring AI requirements integrate with existing frameworks like the Medical Device Regulation and In Vitro Diagnostic Regulation in Europe.

What Does This Governance Divide Mean for Global AI Development?

The split between the US and Europe reflects fundamentally different philosophies about how to balance innovation and safety. The US approach assumes that existing laws and regulatory agencies can handle AI risks without creating new bureaucracies. The European and health-sector approach assumes that AI is different enough to warrant new, specialized oversight mechanisms.

This disagreement matters because policy frameworks that emerge from forums like the G20 Innovation Ministerial could dramatically shape the limits on the power that big tech and AI companies have globally. Companies operating internationally will need to navigate both approaches, potentially building different versions of their systems for different markets.

The stakes are particularly high in healthcare, where AI systems are making decisions that affect human health and privacy. The convergence of US regulators around core tensions between guardrails and innovation suggests that some middle ground may eventually emerge, but for now, the world is watching two competing visions of AI governance unfold in parallel.