OpenAI, the Mansion House Speech, and the Future of Health Governance

The past week has offered a stark study in contrasts for the future of the digital economy.

On one side of the Atlantic, the Mayor of London, Sadiq Khan, used his Mansion House speech to issue a warning: AI could become a “weapon of mass destruction of jobs” unless we actively seize it as a “superpower for positive transformation.”

On the other side, OpenAI effectively launched ChatGPT Health.

For those of us operating at the intersection of AI governance and health, these are not separate news cycles. They are the twin engines of what I call the “Twin Disruption”—a convergence of economic anxiety and technological ambition that is reshaping the health industry faster than regulation can keep pace.

Why health is the new platform play

OpenAI’s announcement is significant not just for its features, but for its direction. By integrating with b.well to access medical records and connecting with wearables like Apple Health, OpenAI is moving to capture the most high-trust vertical in a user’s life.

This mirrors a broader trend we see across the tech sector, similar to Amazon’s moves with One Medical. The objective is clear: maximize the user’s capabilities within the platform. If a patient can review lab results, prepare for appointments, and track fitness trends within a single interface, the platform becomes “sticky.” It creates a cognitive ecosystem that is hard to exit.

But while Amazon is building a physical and logistical trench (clinics and pills), OpenAI is building a cognitive one (data context and reasoning). Both strategies, however, face the same formidable barrier: Trust.

Navigating this barrier requires “calibrated trust.” As AI transitions from a simple tool to an intermediary for our health, the gap between a user’s confidence and the system’s actual limits becomes a critical risk. It is an important trade-off. On one side, under-trust breaks the platform play; if users are too skeptical, the ecosystem loses the ‘stickiness’ and efficiency that make it valuable. On the other, over-trust invites danger; users may inadvertently compromise their privacy or outsource safety-critical decisions to an imperfect machine. Success lies in balancing high utility against the real risks of manipulation and recognizing that while younger generations may be more willing to trade their data for this convenience, that trust is transactional, not unconditional.

The hypothesis: Connectivity without equity is failure

The launch of ChatGPT Health validates a a disruption within the industry we’re seeing: patients armed with AI wearables and ChatGPT are forcing a bottom-up health revolution that stakeholders are just not prepared for.

It is also important to note that while many tools and platforms roll out first in the US, many remain unavailable to users in the UK, the EEA and other parts of the world. This fragmentation illustrates the “equity gap” that often accompanies digital health advancements. While the technology promises to democratize information, like helping 230 million weekly users navigate complex systems, its availability is strictly defined by borders and regulatory regimes. A stark example is ChatGPT Health itself: while US and global users can now join the waitlist, this feature remains geoblocked in the UK and Europe. This delay mirrors the rollout of Apple’s ECG App back in 2018, which arrived in Europe and the UK months after the US due to regulatory validation.

As the Mayor of London highlighted, the anxiety surrounding AI is often about who gets left behind, whether that is the workforce facing displacement or patients navigating a two-tier system.

The governance opportunity

The disconnect between the regulatory caution of certain governments and the commercial speed of major tech companies and the US government, which operates under a sector-specific approach (allowing it to move faster), is a critical challenge for the global industry. At Access Partnership, we see three specific areas where “anticipatory governance” and policy positioning is required to bridge this gap:

1. The Interoperability Trap: Scaling a product like ChatGPT Health requires navigating the EU Health Data Space and the UK’s specific NHS transformation agenda on top of data privacy, AI and platform rules. Success depends on building compliance strategies that can adapt to the fragmentation of rules we’re seeing today.
2. The “Human-in-the-Loop” Mandate: To appease fears of “job destruction,” tech developers must design systems that augment clinicians rather than bypass them. OpenAI’s “HealthBench” and their insistence that the tool is designed to “support, not replace” is a good start, but regulatory and third-party validation of these claims is the next step.
3. Trust as Currency: Trust is the only currency that matters in health AI. A single data breach or misdiagnosis in a “secure” environment could set the industry back a decade. We need transparent safety protocols that go beyond standard “terms of service,” something we are actively looking to help providers and insurers build.

Moving forward

The “Digital Front Door” to healthcare is being built, but it will not open for everyone simultaneously without a deliberate policy strategy.

The winners in this space will be the companies who can navigate or, better yet, shape the complex intersection of patient safety, data privacy, and regulatory fragmentation. At Access Partnership, we are focused on exactly this intersection, helping organizations move from reactive compliance to proactive leadership in the new health economy.