What does a “fair” partnership look like between hospitals and AI developers? As hospital workflows, clinician expertise, and rich clinical data increasingly drive the performance and market value of vendors’ AI tools, providers face important questions. It’s about how we can share the benefits while protecting patients, maintaining trust, and avoiding new forms of liability.
This article examines new models for data sharing, governance, and value capture, providing health systems with a practical framework to ensure they benefit from data-driven innovation.
Hospitals introducing AI
In recent years, more hospitals have developed and licensed AI tools to reduce clinician burnout, improve clinical care, and streamline operations. The tools developed range from those created entirely in-house to those developed in partnership with emerging and established AI companies.
As healthcare costs increase, hospitals are considering implementing AI tools as a cost-saving measure. Specifically, the adoption of AI tools by hospitals is projected to save $900 million in hospital care costs by 2050. Given the cost-saving and efficiency benefits, hospitals have increased their investment and deployment of AI tools at a rapid pace in recent years.
Existing data sharing model
Hospitals utilize large amounts of data on a daily basis and are engaging in large-scale data sharing collaborations to ensure continuity of care.
Health Information Exchange (“HIE”) allows the exchange of health information between organizations within a region, community, or hospital system. HIEs have been developed by both private and public organizations to facilitate rapid information exchange.
Hospitals also rely on the expertise of large technology companies to facilitate information exchange. In June 2025, CMS hosted an event with technology companies, including leading cloud service providers and major search and advertising platforms, to begin laying the foundation for a digital health ecosystem to improve patient outcomes.
In January of this year, a local health system moved its entire technology infrastructure to the cloud using a leading cloud computing platform (Editor’s note: This organization is a customer of the author and/or the author’s company). Additionally, hospitals are beginning to monetize de-identified data, as commercial use of properly de-identified data does not require patient consent. Pharmaceutical companies, in particular, have recognized that de-identified data can help determine how to target clinical trials and refine marketing strategies, creating a large and lucrative secondary market for hospitals to sell de-identified data.
Changes in the industry due to the advancement and spread of AI
Recently, the hospital industry has been transitioning from episodic data transfers to AI-powered continuous data systems to enable proactive patient care. This transition integrates continuous monitoring through items such as wearables and smart beds, replaces regular checks by nurses with command centers, and facilitates operational efficiency and seamless data sharing. The advantage of continuous monitoring is that it provides real-time measurements of the patient’s vitals. This saves costs and reduces the risk of missing signs of deterioration compared to testing patients intermittently.
Additionally, hospitals and healthcare providers are shifting their priorities from one-time innovation projects to embedded AI enterprise ecosystems. Major technology companies (Editor’s note: This company is a customer of the author and/or the author’s company) has launched Agent Health Assistant, developed by its primary care subsidiary. The system works with health information exchanges to provide personalized triage insights derived from a patient’s medical history. Large academic medical system (Editor’s note: This company is a customer of the author and/or the author’s company) After user research showed a 61% reduction in cognitive load and a 77% increase in job satisfaction, we expanded deployment of our enterprise AI platform to more than 2,800 clinicians across 25 hospitals.
Given the positive data regarding enterprise adoption of AI systems, hospitals across the country are transitioning to enterprise AI adoption.
Legal and business considerations
As hospitals pursue partnerships with AI developers, there are several legal and business considerations to keep in mind.
When hospitals form joint ventures with for-profit AI companies, they must navigate complex tax issues to maintain tax-exempt status. This challenge is not unique to AI, but the fast pace of these partnerships means tax considerations are often overlooked. Additionally, hospitals must remain vigilant about HIPPA compliance and maintaining patient trust. Partnerships involving the sharing of clinical data with technology companies carry inherent reputational risks, as illustrated by previous class actions alleging improper disclosure of patient records to commercial partners.
To alleviate these concerns, hospitals should ensure that data used for commercial purposes is appropriately anonymized in accordance with HIPAA safe harbors or professional judgment methods. Parties to a data sharing collaboration should address at the beginning of the partnership whether the hospital, the AI developer, or both will retain ownership of and access to the underlying data and derived datasets generated through the collaboration, and clearly delineate data rights. Additionally, an AI governance framework is essential to ensure that deployed models are clinically valid, unbiased, and subject to continuous human oversight. Finally, hospitals must prioritize risk mitigation through carefully negotiated contract provisions, including comprehensive indemnification clauses, data security representations and warranties, and clearly defined allocation of responsibility to protect the facility in the event of model failure, data breach, or regulatory enforcement action.
Approach AI with strategic rigor
As hospitals rapidly invest in and deploy AI tools, the question is no longer whether these institutions will leverage them, but how they will leverage these tools. Hospitals have vast repositories of clinical data, deep clinical expertise, and established patient relationships that can be refined with AI technology.
To avoid legal and regulatory risks, health systems must approach AI collaboration with the same rigor they apply to other strategic transactions. This means clearly delineating data ownership, ensuring robust HIPAA compliance and patient trust, implementing a comprehensive AI governance framework, and negotiating contractual protections that appropriately allocate risk. In doing so, hospitals can ensure both clinical benefits and cost savings through AI collaboration.
Carolyn Metnick is a partner at Mullin Sheppard and a member of the Healthcare, Privacy and Security team. Timothy Rozier-Byrd is an employee of Mullin Sheppard and a member of the Healthcare Industry team.
