Can AI companies make personal health records work this time around?

More than a decade after Microsoft and Google launched their ill-fated forays into the world of personal health records (PHRs), the Big Tech sector is at it again – this time with a powerful new weapon on their side. 

This week, both OpenAI and Anthropic pushed out a flurry of major announcements around their healthcare goals, with the two prominent AI companies taking aim at how consumers interact with their personal health data. 

First, OpenAI unveiled ChatGPT Health, which allows users to connect health data sources and ask health-related questions in a dedicated environment strengthened with additional privacy features. The company quickly followed up with HIPAA-ready products designed for the provider side, and completed its headline-making trifecta with the$60 million acquisition of Torch, a startup focused on building a “unified medical memory” for AI to connect disparate medical records and improve interoperability.  

Not to be outdone, rival AI powerhouse Anthropic introduced a suite of healthcare capabilities in Claude, its primary AI offering. In addition to support for provider-facing administrative tasks and a set of tools for life science companies, Anthropic is hoping to woo consumers with the ability to give Claude access to their medical records, lab data, and wellness datasets.  

“When connected, Claude can summarize users’ medical history, explain test results in plain language, detect patterns across fitness and health metrics, and prepare questions for appointments,” Anthropic says. “The aim is to make patients’ conversations with doctors more productive, and to help users stay well-informed about their health.” 

The twin announcements set up a fresh new rivalry for dominance in the PHR space, a promising but perilous frontier for tech companies, which typically trade consumer data to advertisers to make a buck.  

Health IT old timers remember the first go-around from Big Tech, when Microsoft HealthVault and Google Health tried and failed to solve the interoperability problem even as it was developing during the glory days of the EHR gold rush. The double withdrawal from the segment seemed to turn PHRs into a no-fly zone, particularly for companies that weren’t native to healthcare and didn’t fully understand the inherent complexity of the task.  

Those attempts are still casting a shadow over the PHR landscape. The resurgence in interest, now with an AI twist, brings up plenty of the same pressing questions, along with some new ones.  

How will consumers use these capabilities to make decisions about their health? How will providers respond? And how will these AI companies, who aren’t projected to be profitable for several years yet, benefit from creating these novel health data ecosystems? 

Explore how healthcare leaders are applying AI to real clinical, operational, and workforce challenges at ViVE. AI Zone @ViVE 2026.

A mix of caution and celebration from the health IT community

Anecdotally, industry voices seem split on whether or not this new iteration of PHRs is a good thing.  LinkedIn and X are abuzz with hot takes on the issue (many written with the aid of ChatGPT, incidentally), with some observers hailing the announcements as the missing link between patients and the ability to take a self-actualized role in their own care management, and others questioning the privacy, security, ethics, and ultimate impact of forking over medical records to AI models that are still prone to hallucinations and incorrect results. 

Many pointed out that consumers are already using ChatGPT and similar tools to help interpret health information, get feedback on worrying symptoms, or find the right providers. OpenAI says that more than 230 million people globally ask health and wellness questions on ChatGPT every week.  

Creating a more secure place to do so – and adding in their personal health data to provide more accurate assessments of what needs attention and what’s probably fine – could just make this established trend safer and more effective for patients. Conversations with providers could get to the point faster and be more productive, while shared decision-making could become a reality for people who have already read up on their options. 

On the other hand, some pundits, particularly those with clinical backgrounds, were quick to note that ChatGPT and Claud provide little transparency into their decision-making processes and are notorious for making errors, some of which might be serious. There is no clinical oversight built into these platforms, and patients could be making decisions without real clinical input that might result in some form of harm. 

Still more expressed concern about the ongoing trend of internet medical sources eroding the trust between patients and their care providers. In a matchup between ChatGPT and a human doctor, how often will patients feel like omnipresent, omniscient AI has the better answers and advice? 

And then there are the inevitable straight-up techie concerns, including issues about connecting records from traditional healthcare organizations and validating the correctness and completeness of the data before AI platforms go to work on them. Others predicted that claims of souped up privacy and security protections – as well as promises that personal health data won’t be used to train models – are a bunch of smoke and mirrors in an industry where breaches are inevitable and people are the products. 

Following the money: Shifting incentives may give hope to PHR proponents

Interoperability isn’t profitable in healthcare, which is one of many reasons why earlier attempts to create PHRs never took off. Data siloes might be bad for patients, but they help fee-for-service healthcare organizations keep control of their populations, which has long been a pain point in the fight to free up access to personal data. 

Freely available AI isn’t profitable, either. While Anthropic is slated to start making real money a lot faster than OpenAI, neither company can afford to offer these enhanced healthcare capabilities to the marketplace out of the goodness of their hearts. 

Simply selling personal health data, even with claims of full deidentification, would be a radioactive red flag for either company. So why invest in PHRs?  

I asked ChatGPT to explain its own creator’s thinking, and the answers made pretty good sense. 

“ChatGPT for Health is not a consumer health app looking for revenue. It’s a distribution beachhead,” I was told. “OpenAI’s incentives align with being the operating system for health intelligence, not a healthcare company.” 

The money is partly in premium consumer subscriptions to higher-end health data management features, it said, but primarily in enterprise licensing to healthcare systems and integration partnerships with healthcare technology companies that want to build on ChatGPT for Health’s foundational functions. 

By monetizing dual roles as a) the trusted, in-demand mouth of the funnel for health data as patients aggregate disparate records, ask questions, and clarify their care priorities, and b) the connective tissue across the health system distributing access to data, intelligence, AI capabilities, and insights that narrow the gap between patients, providers, and payers, companies like OpenAI and Anthropic may be able to build the smart, AI-enabled interoperability infrastructure and data pipelines we’ve all been searching for. 

This strategy shifts the incentives of interoperability – although the main rewards will flow to the folks in control of the infrastructure. If companies really can start making money off of moving data, which they couldn’t do ten years ago when the health IT landscape was too immature and consumers weren’t used to thinking about their health in that way, then there’s a hope that PHRs really could stick. 

Idealistic dreamers might even wonder if the downstream effects could include a reinvestment in value-based care, now that financial strategies that rely on data guarding start to break down, patients can become more engaged more easily, and consumers renew their demands for more preventive care and faster, more comprehensive answers to their problems.  It might be a longshot, but it’s nice to think positively about the possibilities.  

Finding the right path forward into AI-enabled PHRs

Success for either company depends on a wide range of factors, including uptake from enough consumers to make it worthwhile for enterprise investment. Just being big and well-known with a large built-in user base is no guarantee of long-term viability, as the ghosts of PHRs past have illustrated.  

The key will be how these companies (and any future rivals) monetize and incentivize engagement from all the nodes of the healthcare data spiderweb – and how well they build and maintain trust among stakeholders who are letting them into one of the most sensitive and high-risk areas of the tech world. 

If consumers bite at the chance to trade in Dr. Google for Dr. ChatGPT and Dr. Claude, and providers find AI-informed patients helpful rather than headache-inducing, we might see a sea change in patient engagement that could catalyze the next era of value-based thinking.  

But if either company is hit with a major privacy and security scandal, they get overeager and price themselves out of contention with enterprise clients, or inertia from the industry leaves them bogged down in technical difficulties, then PHRs could fizzle out once again. 

It will be fascinating to see how these dynamics will unfold over the next several years, and what new products – and pricing structures – AI companies will develop to further mine the potential riches that await in the healthcare industry. 

Jennifer Bresnick is a journalist and freelance content creator with a decade of experience in the health IT industry.  Her work has focused on leveraging innovative technology tools to create value, improve health equity, and achieve the promises of the learning health system.  She can be reached at [email protected].