In Simple Terms
Epic says it wants to make it easier for other companies to use information stored in its medical-record systems to build AI tools. But critics say some app developers may not get equal access. The challenge is to open the system to more users without making it harder for competitors to build and connect their products.
A Debate Over Who Gets Access
At a discussion on Wednesday, November 18, 2026, during the HLTH health technology conference in Las Vegas, executives from Epic Systems Corp. faced a question that went beyond software details: Who can use the medical information stored in the company’s systems, and how? Epic said it is working to make it easier for companies to use AI to read data held in its systems, while defending itself against accusations about competition and access for developers of competing apps.
Epic says it wants to make it easier for other companies to use information stored in its systems to build AI tools. But the company is also facing claims that it does not give all app developers the same access to that data. As Epic seeks to open up its system while receiving a large number of requests, a central question remains: How can access be expanded without creating an obstacle for competitors?
From Medical Records to AI Tools
Epic works in electronic health records: digital systems that store medical information within healthcare organizations. According to information its executives shared during the discussion, its platforms contain records for more than 280 million Americans. That figure shows the scale of the information at the center of the debate, but it does not, on its own, tell us what any outside company can see or which kinds of data a particular app can use.
Epic says its aim is to make it easier for AI companies to use data held in its platforms. As described during the discussion, that means making it easier for companies to access data or connect their software to Epic’s systems—not giving every company open access to every record. An announcement about making data easier to use does not, by itself, spell out the conditions for access or explain what approvals and limits apply to each app.
This issue matters because AI software needs information to do its job. But the fact that data is stored in a healthcare system does not mean that transferring or reading it is simple. An app developer needs a suitable way to connect to the system holding the records, and its needs may differ from those of another company. Connecting systems is therefore both a practical and a commercial question: How can those connections be made, and who decides what access looks like?
75,000 Inquiries in One Year
Seth Howard, Epic’s executive vice president of research and development, offered a figure that illustrates the pressure on the company: It received about 75,000 inquiries in the previous year from organizations seeking access to its data systems. The number highlights the volume of requests Epic says it faces, but it does not necessarily mean that every inquiry came from an independent app developer or led to a completed connection project. An inquiry may be the start of a request; it is not proof that access was granted or that a system was integrated.
Howard said that building a custom connection for every organization one by one would leave the company spending most of its effort creating those connections. His point can be understood as a practical argument: When requests become numerous and varied, handling each one separately can turn into a major burden. But that does not resolve the opposing question: Do companies have clear, common routes for integrating with Epic’s systems, and does the way requests are handled affect developers’ opportunities?
This is the trade-off Epic is trying to explain. On one hand, the company points to the volume of requests and the difficulty of designing a separate connection for each organization. On the other, app developers need to know what steps they can take, what requirements apply, and whether they have roughly equal opportunities. A high number of requests helps explain why the task is complicated, but it does not, by itself, answer whether access is fair or sufficient.
Accusations of Favoritism, and a Defense of Openness
The discussion followed an investigation by STAT into how Epic makes data available to developers of competing apps. According to the article, the investigation found that the company selectively withholds access to data some developers need, weakening competition. Epic executives responded to that criticism directly for the first time during the event. They said the company is working to build a more open environment while receiving a large number of requests from outside software providers seeking to connect to its systems.
The accusation and the company’s response should be considered separately. The allegation is that selective access can harm developers of competing apps; Epic, meanwhile, says it wants to make its system more open and describes the volume of integration requests as an operational challenge. The information available about the discussion does not settle which account is more accurate, nor does it provide enough detail about specific decisions involving particular companies or apps. The debate is therefore about how access rules are applied, not simply about whether Epic has announced a desire to be more open.
The issue is sensitive because a company that manages a large platform can, in practice, influence whether other software can work with the data it holds. If the steps for connecting are clear and easy to understand, developers can assess what they need to do. If the standards or decisions are unclear, they may wonder whether some requests receive different treatment. These questions arise from the nature of the dispute, but the article about the discussion does not establish them as proven findings.
Why the Dispute Matters Beyond Software Companies
At first glance, the disagreement may seem to concern only companies that make apps. But it also affects how digital tools might gain access to health information. Any app that relies on data stored in another system needs a connection between the two. If that connection becomes easier to establish, companies may have more opportunities to build tools that use medical records. If integration remains dependent on separate, complicated processes, developing those tools could become harder or slower.
Making connections easier, however, is not enough to establish the value of any AI tool, nor does it prove that an app will improve care or produce better results. The announcement at the event concerned access to data and connections to Epic’s platforms, not clinical outcomes or trials demonstrating the benefit of a particular product. It is important not to confuse enabling companies to read data with proving that every new use is successful or appropriate.
The published information also does not explain whether the promised changes will alter how every company gets access, when they might happen, or what safeguards will come with them. These details matter because the phrase "easier access" could describe very different arrangements: clearer connection procedures, less need to build a custom connection for each organization, or other routes that were not explained during the discussion. Without implementation details, the phrase describes a stated direction rather than a complete account of a specific process.
Connecting Systems—and Defining Data Limits
Creating a link between an outside app and a medical-record system does not necessarily mean the app gets everything in that system. Access may mean that a program can request or exchange a specific type of information, rather than getting access to the entire database. Since the discussion did not provide a list of data types or app permissions, it is not possible to say exactly what AI tools will be able to read after any change Epic announces.
The large number of requests also raises questions about how priorities are set. Are requests handled according to consistent rules? Do timelines or requirements vary depending on the type of service? The article does not answer these questions. But they are reasonable things to ask of any organization handling tens of thousands of inquiries, particularly when it manages a platform containing records on a vast scale. Answers may be necessary to assess whether the connection process can be expanded while remaining fair.
Epic’s explanation focuses on the effort required to create individual connections. The company says that responding to every organization with a custom connection would consume its staff’s time. This points to an important distinction between the volume of requests and the ability to meet them: Demand for integration may be high, but responding at scale requires a repeatable approach. Even so, the number alone does not reveal how many requests were approved, rejected, or remain under review, or how long they took to handle.
What to Watch After the Announcement
The next step in assessing Epic’s position is not simply to take the company’s stated desire for a more open environment at face value, but to look at what it does in practice. It will be useful to know whether Epic publishes clear rules for connecting to its systems, whether those rules are equally available to different app developers, and how response times for requests can be measured. These details were not provided at the event, but they would help show whether the announced direction leads to tangible change.
It is also important to distinguish among three things that often get blurred together in this debate: a company holding records in its platform, an outside program being able to connect to that platform, and an AI tool being able to use the information made available through that connection. The announcement addressed making it easier for companies to use data in Epic’s systems, but it did not fully describe every step in that chain. More details about the conditions and scope of connections would make it possible to identify what is actually changing.
The competition-related accusations also require an assessment that goes beyond general statements from either side. A company’s claim that it is working toward openness does not, by itself, prove that all developers have equal opportunities. Likewise, the allegation in the investigation does not, on its own, resolve the details of every individual case. The discussion brought into public view a clash between criticism that access is being selectively withheld and Epic’s defense, which emphasizes the challenge of handling the volume of requests and its ambition to build a more open ecosystem.
The Bottom Line
Epic’s comments have put access to medical records at the heart of a debate about AI in healthcare. The company says it is working to make data in its platforms easier to read, while its approach faces criticism over whether developers of competing apps can access that data. Seth Howard cited about 75,000 inquiries the company received in the previous year, while Epic’s systems hold records for more than 280 million Americans.
Those figures and promises are not enough to show whether access will improve or competition will become more equal. The available information leaves key details unclear, including how the changes will be implemented, what standards will govern requests, and what data will be available. The announcement’s significance will depend on the practical mechanisms that emerge: understandable rules, workable connection routes, and treatment that can be assessed for consistency. Until then, the challenge remains twofold: handling thousands of requests without building every connection from scratch, while addressing developers’ concerns about access and fairness.