Electronic Health Record (EHR) companies have historically used proprietary formats to lock in customers. AI's ability to read and translate unstructured data from any source effectively breaks these data silos, finally making patient data truly portable.

Healthcare has historically been a service, with costs tied to licensed professionals. AI models like Gemini and ChatGPT are changing this by providing medical advice, effectively turning healthcare into a product. This shift, currently tolerated by regulators, could dramatically lower costs and increase access, just like software products.

We possess millions of data points on interventions, but they are useless to AI models because they're trapped in thousands of disparate EMRs in varied formats. The challenge is not generating more data, but solving the human incentive and alignment problems required to create unified data registries.

Despite processing 15 million clinical charts, Datycs doesn't use this data for model training. Their agreements explicitly respect that data belongs to the patient and the client—an ethical choice that prevents them from building large, aggregated language models from customer data.

Jurgi Camblong argues the concept of "data ownership" is a misnomer, citing GDPR's framework of data subjects, controllers (hospitals), and processors. By positioning as a trusted processor rather than an owner, Sophia Genetics gains access to a wider, decentralized network, which is more powerful than a centralized, owned dataset.

As AI allows any patient to generate well-reasoned, personalized treatment plans, the medical system will face pressure to evolve beyond rigid standards. This will necessitate reforms around liability, data access, and a patient's "right to try" non-standard treatments that are demonstrably well-researched via AI.

Claire Smith envisions a new biotech business model focused on aggregating vast, unstructured health data (genomic, clinical notes) to sell high-value insights to pharma. This "Palantir-style" approach turns data into a scalable product for target identification or patient stratification, avoiding the traditional drug development path.

To overcome the scaling challenges of traditional biobanks, Regeneron is pioneering a new model. They partner with companies specializing in aggregating de-identified health records and, separately, with groups handling bio-sampling. This "uncoupled" approach allows them to link massive, independent data streams to achieve unprecedented scale.

While hospitals and insurers are bound by HIPAA, their terms of service often include clauses allowing them to sell de-identified patient data. This creates a massive, legal shadow market for healthcare data. AI companies will leverage this data, obtained via consumer consent, to build powerful advertising and personalization engines.