Creating a Healthcare-Focused Artificial Intelligence Public-Private Platform for New York State
- Sally Dreslin
- 15 hours ago
- 18 min read
Updated: 9 minutes ago
Enhancing Access to AI Tools for Healthcare Providers Across New York
PDF version:
Key Takeaways
This paper proposes establishing a healthcare-focused, AI public-private platform to enable broader access to AI’s benefits in patient care, outcomes, and operational efficiency, and to help close the technological divide among healthcare entities.
New York State has two important precedents for a new public-private platform involving health information technology and AI: the State Health Information Network of New York, known as “the SHIN-NY,” and Empire AI.
There is increased urgency to improve the outcomes and efficiency of New York’s health delivery system following recent federal legislation that is expected to further compromise the fiscal condition of the healthcare industry, particularly for distressed hospitals.
The public-private utility model leverages private-sector expertise and investments that have already produced robust AI tools and aligns them with the social responsibility of expanding access.
Developing and maintaining AI infrastructure is resource-intensive. By creating a statewide utility, private partners can contribute technical knowledge and infrastructure, while public oversight ensures that access is fair, secure, and accountable.
Incorporating AI tools into healthcare delivery is not without concerns and challenges. In any context, the responsible use of AI technologies requires a variety of considerations, and in healthcare, the stakes can be even greater.
Tapping into the expertise of New York’s Empire AI consortium and collaborating with private partners who have expertise in building private AI environments for healthcare, the State could create a secure, private-cloud environment that leverages existing AI technologies, such as large language models and other tools.
As healthcare entities and services consolidate and financial pressures intensify, supporting smaller hospital systems and community-based healthcare entities in leveraging statewide resources through the creation of a healthcare-focused AI public-private utility will support the sustainability of those providers and help them access technology and tools that would otherwise be out of reach.
Introduction
This paper explores the idea of a shared artificial intelligence (AI) platform, offered through a public-private partnership, that could at least partially bridge the gap between the “haves” and “have-nots” of healthcare providers when it comes to the deployment of AI technology in the sector.
AI has the potential to enable significant changes in the way healthcare is delivered, improving patient satisfaction, healthcare outcomes, and provider operating efficiency. As with electronic health records (EHR) and health information technology (HIT) generally, the benefits of this technology will be constrained if important parts of the healthcare delivery system are left out because they lack financial resources and/or management capacity.
New York State (NYS) has two important precedents for a new public-private platform involving health information technology and AI. First, the State Health Information Network of New York (known as “the SHIN-NY”) offers all providers the ability to be connected in a health information exchange (HIE) in which clinical patient data can be securely transferred in a regulation-compliant environment.[i]
Second, Empire AI is an NYS-sponsored consortium that includes Columbia University, Cornell University, CUNY, New York University, Rensselaer Polytechnic Institute, SUNY, the Simons Foundation, University of Rochester, Rochester Institute of Technology, and the Icahn School of Medicine at Mount Sinai. By providing access to high-performance computing resources funded by New York State and private philanthropists, which otherwise would be limited to large private corporations, Empire AI enables academic institutions to focus on AI’s unique capabilities to solve complex problems.[ii] Both the SHIN-NY and Empire AI are non-governmental non-profits that work in close cooperation with NYS.
The challenges of democratizing the benefits of AI will confront governments at all levels across the country. New York has the opportunity to be a pioneer in developing a public-private partnership that ensures the benefits of AI are made available to as broad a healthcare sector community as possible, while preserving the ability of institutions such as academic medical centers to push the envelope of the functionality of this evolving technology.
The urgency for improving the outcomes and efficiency of New York’s health delivery system has been heightened by the enactment at the federal level of the H.R. 1 tax and spending reconciliation law on July 4, 2025. New York is facing an convergence of crises in its healthcare delivery system: federal changes that could result in the loss of billions of dollars to the system; loss of health insurance coverage for an estimated one and a half million New Yorkers; loss of supplemental food benefits; potential loss of housing support (food and housing are important social drivers of health); a significant existing and growing health workforce shortage that will likely be worsened by the recent changes in federal immigration policy; and a strained primary care and rural health infrastructure, among others.
New York’s hospital associations have estimated that the new federal provisions will cost hospitals approximately $6 billion directly and up to $8 billion in related impacts, once fully implemented. As the Step Two Policy Project has discussed previously, this loss represents,
“… roughly 7% of New York’s total hospital revenue of approximately $120 billion, a staggering decline in operating margin given that nearly 60% of hospitals in New York State already operate at flat or negative operating margins.” Without the State providing operating subsidies to make up for at least some of this lost revenue, New York’s hospital infrastructure may look very different when the dust settles.”
Given the serious financial risk to New York’s healthcare delivery system, the risk of inequities in access to AI-driven innovations and efficiencies is very real. The major EHR vendors are expanding the incorporation of AI tools into their EHR offerings,[iii],[iv] further widening the gap between the “haves” and the “have nots,” who may be using less sophisticated EHR programs or those still dependent on basic systems or even paper records. Healthcare entities will need to enhance their ability to provide efficient, effective, high-quality care and to streamline and modernize their operations to remain sustainable in an increasingly difficult fiscal environment. AI tools can play an important role in these efforts, improving administrative functions and operations.
Supporting Rural Healthcare
NYS is aggressively working to expand access to mobile and broadband service across the state through the ConnectALL program. In the Step Two Policy Project’s Healthcare in Rural New York: Current Challenges and Solutions for Improving Outcomes from October 2024, we described the challenge of accessing broadband service in rural NY, noting that, “many rural counties will continue to face structural barriers to realizing the full potential of telehealth” and other benefits to the health delivery infrastructure from connectivity.
The recent H.R. 1 tax and spending reconciliation bill includes a provision called the Rural Health Transformation (RHT) Program. In the funding announcement, CMS states that the program is intended to promote “innovation, strategic partnerships, infrastructure development, and workforce investment,” and identifies five strategic goals. Three of these strategic goals could be supported through the creation of the EmpireHealthAI. They are:
“Sustainable access: Help rural providers become long-term access points for care by improving efficiency and sustainability. With RHT Program support, rural facilities work together—or with high-quality regional systems—to share or coordinate operations, technology, primary and specialty care, and emergency services.
“Innovative care: Spark the growth of innovative care models to improve health outcomes, coordinate care, and promote flexible care arrangements. Develop and implement payment mechanisms incentivizing providers or Accountable Care Organizations (ACOs) to reduce health care costs, improve quality of care, and shift care to lower cost settings.
“Tech innovation: Foster use of innovative technologies that promote efficient care delivery, data security, and access to digital health tools by rural facilities, providers, and patients. Projects support access to remote care, improve data sharing, strengthen cybersecurity, and invest in emerging technologies.”
Artificial Intelligence in Healthcare
The rapid expansion of the use of artificial intelligence across industries is transforming operations, decision-making, and user experiences, with applications ranging from automation and robotics to generative models such as ChatGPT, Gemini, and Claude. In healthcare, AI is being used to interpret diagnostic imaging and EKGs, support triage and clinical decision-making, monitor patients remotely, develop personalized care plans, predict disease risk, summarize and visualize data, and streamline administrative work such as clinical documentation, prior authorization, and supply chain and revenue cycle management. Additionally, patients are accessing open AI tools to self-diagnose, treat, and seek second opinions.
However, with these opportunities come some significant challenges and risks. Concerns about patient privacy, data security, algorithmic bias, and errors in automated decision-making highlight the need for oversight, transparency, and rigorous validation to ensure patient safety and trust.[v]
These concerns are particularly acute when it comes to compliance with the Health Insurance Portability and Accountability Act (HIPAA).[vi] For example, entering identifiable protected health information (PHI) into open generative AI tools like ChatGPT, Gemini, or Copilot is not HIPAA-compliant unless an appropriate business associate agreement exists between the healthcare provider and the AI tool,[vii] often accomplished through an enterprise licensing agreement.
This constraint disproportionately affects smaller healthcare organizations, such as safety-net, rural and critical access, or sole community hospitals; community clinics; emergency medical services (EMS), and other providers that may lack the financial or technical resources to implement secure AI functionality or benefit from AI-enabled electronic health records used by larger, better-resourced institutions. There are numerous HIPAA-compliant, healthcare-focused applications and platforms available to the healthcare sector, but they can be costly to purchase, implement, and monitor.
Implementing responsible AI in hospitals and other clinical provider settings requires significant financial investment, both upfront and ongoing. Healthcare entities must allocate funds for acquiring enterprise-wide AI platforms, which may include licensing fees, subscription costs, or expenses related to integration with existing EHR and other systems. Beyond the technology itself, additional costs arise from data storage and security infrastructure, which must meet regulatory standards. Healthcare entities also need to budget for staff training, clinical workflow redesign, and governance structures such as ethics boards or compliance reviews to ensure safe use.[viii]
In addition to financial commitments, healthcare entities need technical resources to deploy AI tools responsibly. This includes secure, high-capacity data infrastructure capable of handling large volumes of clinical and operational data, as well as sufficient cybersecurity to safeguard protected health information. Data scientists and clinical IT specialists must provide technical expertise and collaborate with medical, pharmacy, and other staff to validate algorithms, monitor performance, and address issues such as accuracy and bias. Interoperability with existing clinical systems, including EHRs, order entry systems, and diagnostic imaging networks, requires technical skill and ongoing maintenance. Additionally, entities must establish continuous monitoring and feedback mechanisms to ensure AI tools are not only accurate when first implemented but also remain reliable and clinically relevant as patient populations, medical evidence, and best practices evolve.[ix]
These significant investment requirements and ongoing commitments to achieve the benefits of AI make financial resources and sustainability significant concerns. While larger academic medical centers and health systems may have access to capital or external partnerships, lesser-resourced providers such as safety-net, rural and critical access, or sole community hospitals; community clinics; nursing homes; or EMS agencies often lack the resources to purchase or maintain AI technology. This gap in resources between well-resourced providers and other critical players in the healthcare delivery system threatens to lead to inequities in the quality of care as well as in access to the benefits of AI-driven innovations and efficiencies in healthcare delivery.[x]
Rationale for Creating a Shared Public-Private Healthcare AI Partnership
The concept of a shared platform offered through a public-private partnership is analogous to the decision in 2006 to establish the New York eHealth Collaborative (NYeC) as a public-private partnership with leadership and support from the New York State Department of Health (DOH). NYeC is a non-profit organization with a Board of Directors. NYeC’s mission is to “improve healthcare by collaboratively leading, connecting, & integrating health information exchange across the state,” and the mission of the SHIN-NY, which is operated by NYeC, is to “improve healthcare through the exchange of health information whenever & wherever needed.” The role of the SHIN-NY is to facilitate “the secure electronic exchange of patient health information and connect[s] healthcare professionals statewide.” This health information exchange (HIE) is accomplished in partnership with DOH, with NYeC developing and managing the technology platform that supports the exchange, ensuring that the SHIN-NY provides access to a patient’s data on a statewide basis. NYeC works closely with DOH and receives much of its funding from the State and federal government, although there are no governmental appointees on its Board of Directors.
The belief behind the establishment of NYeC and the SHIN-NY was that maximizing the use of health information technology throughout the healthcare delivery system required secure, interoperable HIE. At the time of its formation and continuing somewhat to this day, the SHIN-NY was criticized by many of the large health systems in New York as being unnecessary and inferior to the proprietary technology platforms operated by those health systems. This criticism, however, misses the main point, which is that an HIE such as the SHIN-NY is essential for statewide program operations and for broadening participation to a diversity of providers and their patients.
The SHIN-NY fills an important gap for provider types that are not included in the closed, proprietary HIT systems of large health systems. The SHIN-NY plays a critical role in enhancing care coordination and public health workflows, including sending admission/discharge/transfer alerts for providers and care management entities, exchanging patient lab results between treating providers, facilitating disease surveillance, and reducing duplicative diagnostic testing. During the COVID-19 pandemic, the SHIN-NY was essential in supporting public health efforts to control virus transmission, for example, by facilitating the sharing of test results. With updated regulations in 2024 that the Step Two Policy Project has written about previously, the DOH and the SHIN-NY are seeking to “establish a framework for modernizing health information exchange throughout the state in further support of SHIN-NY participants and New York’s public health and Medicaid needs.”[xi]
Establishing a shared AI platform supported by a public-private partnership would serve a similar purpose to the SHIN-NY. It could mitigate some of the imbalance of resources between large health systems and smaller providers, including both safety net and community hospitals, community health centers, long-term care facilities, EMS providers, and others.
The exact structure and scope of this public-private partnership “utility” model to make existing healthcare-related AI services more broadly available needs to be thoughtfully planned in concert with all relevant stakeholders. It is unlikely that this platform could provide all the AI-related tools and functionality available to healthcare providers through direct relationships with private vendors. However, it may well be possible to provide a set of core AI tools or services that are compliant with HIPAA and other governmental regulations and otherwise not accessible to some stakeholders.
A HIPAA-Compliant AI Environment for Healthcare Entities
Tapping into the expertise of New York’s Empire AI consortium and collaborating with private partners who have expertise in building private AI environments for healthcare, the State could create a secure, private-cloud environment that leverages existing AI technologies, such as large language models (LLM) and other tools. This private-cloud layer could evolve as new AI models emerge, thus freeing NYS from being reliant on a single set of tools. The State and its partners would ensure this private environment is HIPAA-compliant, aligned with other regulatory requirements regarding the handling of patient data, and accessible only to NYS healthcare entities’ users. The target stakeholders would be NYS-operated healthcare facilities, lower-resourced hospitals such as safety-net, critical access, or sole community hospitals; community health centers; nursing homes; EMS agencies; and other healthcare entities and providers that lack the resources to implement and maintain enterprise-scale AI tools into their work.
Range of AI Functionality in Healthcare
There are a variety of potential uses for AI in healthcare. The applications are in both patient care and in overall administration and management functions. S.M. Varnosfaderani and M. Forouzanfar provide a robust review of The Role of AI in Hospitals and Clinics: Transforming Healthcare in the 21st Century (2024). They offer the figure below as a guide to the areas addressed in their review.
There are many different generative AI tools and technologies available. LLMs, e.g., ChatGPT, Gemini, or Claude, that form the foundation for many AI tools, are described by IBM as,
“… a category of foundation models trained on immense amounts of data making them capable of understanding and generating natural language and other types of content to perform a wide range of tasks. … In a nutshell, LLMs are designed to understand and generate text like a human, in addition to other forms of content, based on the vast amount of data used to train them. They have the ability to infer from context, generate coherent and contextually relevant responses, translate to languages other than English, summarize text, answer questions (general conversation and FAQs) and even assist in creative writing or code generation tasks.”
By leveraging LLMs and other AI tools, the healthcare sector has opportunities to transform disease diagnosis and prognosis, enhance personalized medicine, transform hospital logistics and resource management, automate administrative tasks and optimize patient flow and scheduling, reshape radiology and pathology, integrate wearable devices into continuous patient monitoring, and expand the effectiveness of remote patient engagement via telemedicine.[xii] In hospital management and operations, for example, facilitating access to AI tools to assist healthcare entities across the state in streamlining and modernizing their operations could make progress in leveling the increasingly unbalanced healthcare delivery system in New York by generating savings.[xiii] This could also ultimately help the State reduce its spending on financially distressed hospitals, which reached approximately $3.5 billion for FY 26.[xiv]
The table below, from Varnosfaderani and Forouzanfar’s review, provides an example of some of the potential applications related to operational and back-office functions:
Concerns with AI in Healthcare
Incorporating AI tools into healthcare delivery is not, of course, without concerns and challenges. In any context, the responsible use of AI technologies requires a variety of considerations, and in healthcare, the stakes can be even greater.[xv], [xvi], [xvii], [xviii] Important considerations when incorporating AI tools into healthcare include, among others,
Data Use. Knowing how the system is using personally identifiable, confidential, and sensitive information, and ensuring that its use is ethical and legal.
Human Oversight. Ensuring there is human oversight, i.e., a “human in the loop,” to facilitate accuracy and safety in terms of decision-making that impacts people/patients.
Fairness and Bias. Monitoring training data sets to ensure a diversity of sources and observing and mitigating algorithmic biases.
Transparency and Accountability. Disclosing how the system works, and its capabilities and limitations, ensuring that the use of AI tools is being disclosed, ensuring that the pathway to decisions is traceable, particularly in a situation where there may be an unexpected patient outcome, and making system evaluation results available.
Privacy and Security. Protecting patient privacy and responsible sharing of data; managing consent issues; and ensuring the systems are protected against security breaches.
Cost and Sustainability. The cost of acquiring AI technology can be high, and beyond the tools, infrastructure and human resources are essential for sustainability. Incorporating AI tools into healthcare may necessitate modernizing the existing technology infrastructure.
Varnosfaderani and Forouzanfar’s review provides the following graphic related to navigating the ethical considerations of AI healthcare,
Robust governance and continual monitoring and reporting are critical in addressing these various concerns. Strong governance ensures that artificial intelligence tools in healthcare are developed and used in alignment with ethical values, requirements for patient safety, and regulatory responsibilities. Continual monitoring is essential to detect algorithmic bias or inaccuracies that may emerge over time as data and clinical contexts evolve. Responsible governance and monitoring build trust and accountability, helping healthcare entities maximize the benefits from AI tools while working to minimize the risks.
Why a Shared Platform Model?
The public-private utility model leverages private-sector expertise and investments that have already produced robust AI tools and aligns them with the social responsibility of expanding access. Developing and maintaining AI infrastructure is capital-intensive. By creating a statewide utility, private partners can contribute technical knowledge and infrastructure, while public oversight ensures that access is fair, secure, and accountable.[xix] This structure allows New York’s healthcare system to benefit from private-sector innovation without leaving smaller entities behind.
By integrating public investment, private development, and shared governance, this public-private utility will become a means of democratizing AI access across New York’s healthcare sector. Healthcare entities of all sizes could access the same secure system, gaining access to AI tools without duplicating costs. This approach fosters economies of scale, maximizes efficiency, and ensures that AI in healthcare operates as a shared public resource—strengthening equity, financial resilience, and patient outcomes across New York.
Implementing a Healthcare-Focused AI Environment as a Public-Private Utility
Providing a healthcare-compliant, private-cloud AI platform could involve leveraging the resources of the Empire AI consortium, and could be called, for example, “EmpireHealthAI.” Empire AI was formed to “establish a consortium of public and private research institutions advancing AI research for the public good.”
New York can become a national leader by creating a HIPAA-compliant, secure, private-cloud environment that leverages existing AI technologies, such as large language and other models. With this private-cloud layer evolving as new AI models emerge, NYS would not be reliant on a single vendor or set of tools. By aligning with the Empire AI consortium’s vision and incorporating additional private and healthcare partners, EmpireHealthAI could function as a public-private utility, accessible to healthcare entities across the State. The closed platform would be accessed through an application programming interface (API), as described in the graphic below.
EmpireHealthAI should have a strong relationship – from both a technical and governance perspective – with the Statewide Health Information Network for New York (SHIN-NY). Ideally, an EmpireHealthAI governance board would be established, consisting of EmpireAI consortium members, State and independent healthcare experts, private-sector healthcare AI partners, and SHINY-NY representatives. This Board would be responsible for addressing the concerns discussed earlier related to the use of AI in healthcare. Specifically, setting standards around data use, transparency, and monitoring, including ensuring bias detection, accuracy validation, and ethical use of AI outputs. Safeguards would emphasize patient consent, privacy protections, and alignment with existing regulatory requirements, ensuring the platform supports efficient and effective statewide healthcare delivery, without compromising trust. The Board would also be responsible for deciding on the core AI tools that would be available for healthcare entities to access. Initially, it seems reasonable to offer access through a secure and private-cloud layer to an LLM such as ChatGPT (OpenAI), Gemini (Google), or Claude (Anthropic). The enterprise versions of the open models are expensive, with costs varying based on the use, e.g., for automating processes and building AI agents, for writing code, etc., and are often priced on a per-user and/or usage basis.
From a financial perspective, with EmpireHealthAI facilitating access to existing LLMs and eventually other AI models rather than building from scratch would greatly reduce initial development costs. However, maintaining a healthcare-compliant environment, including ongoing auditing and security protections, will require ongoing funding. A public-private financing structure could sustain the system, with the State subsidizing baseline infrastructure and one-time development of the private-cloud environment. The initial build of the State’s secure AI environment to enable providers to “plug in” without needing to run their own, could be accomplished with an investment in the range of $1 million.
EmpireHealthAI would work together with healthcare entities to evaluate their existing technology and cloud access and connect them to the AI environment. This one-time, per-provider implementation cost to access an LLM could be in the range of $100,000 to $500,000 for the provider entities.[xx] As EmpireHealthAI evolves, however, and adds tools such as AI models for ambient scribing or coding, the per-provider implementation process will become more complex and may require certain infrastructure on the part of the healthcare entities. These processes could be developed through stakeholder consultation. The public-private utility would ensure an affordable, tiered pricing model. Pricing could be tied to organizational size and usage levels, ensuring small providers have access, while larger systems contribute proportionally more. The State could subsidize the costs, ensuring both equity and sustainability while reducing duplicative investments across healthcare entities.
On the infrastructure side, the Empire AI consortium could contribute to the technical backbone of the system, as necessary. The closed nature of the system would ensure that sensitive health data never flows into open commercial AI environments, while still allowing the model to generate value for users. By establishing the platform as a healthcare-focused, public-private utility—anchored to Empire AI, connected to SHIN-NY data, and maintained through a sustainable pricing model, New York could lead the country in building a responsible and secure AI infrastructure for healthcare.
Conclusion
The Health Care Briefing Book for the FY 26 Executive Budget identifies 29% (75 of 261) of NY’s hospitals as financially distressed and states that “many providers are struggling to remain solvent, impacted by many of the same cost pressures that are impacting businesses in other sectors.”[xxi] The State has continued to provide extraordinary financial support to a variety of financially distressed safety net providers. Creating access to leading-edge technology as a shared utility, however, could be an opportunity to fundamentally modernize and streamline operations and patient care for New York’s healthcare entities that are “struggling to remain solvent.” AI tools such as ambient scribes, tools to manage prior authorization, revenue cycle management, and robotic process automation for back-office applications such as billing can save money, improve efficiency, and free staff to focus on patient- and client-facing activities. EmpireHealthAI’s private-cloud environment will also provide an opportunity to thoughtfully and gradually apply AI in clinical areas.[xxii]
This proposal of an AI public-private utility fits Governor Hochul’s strategy of harnessing “emerging technologies for public good.” The Governor’s new pilot to train NYS employees to use AI responsibly embraces the opportunities of AI while appreciating the challenges. Given the profound significance of AI for all aspects of healthcare and the likelihood that many stakeholders will not have the resources to establish their own environment, the State should take the lead in developing a comprehensive public-private partnership to implement this proposal. As the Governor stated in her announcement of the State pilot, “[f]rom the moment we first announced Empire AI, I vowed to put New York State at the forefront of the AI revolution and to ensure that our journey forward was safe, responsible and thoughtful.”
As healthcare entities and services consolidate and financial pressures intensify, smaller hospital systems and community-based healthcare entities need support. By leveraging statewide resources to create a healthcare-focused AI public-private utility, New York can help healthcare providers access technology and tools that would otherwise be out of reach. This will result in improved patient care and increased sustainability, which will enhance access to healthcare for New Yorkers.
Endnotes
[ii] Governor Hochul Announces $90 Million Plan to Expand Historic Empire AI Consortium and Enhance Computing Power for Public Good on Behalf of New Yorkers, February 21, 2025.
[iii] Artificial Intelligence, Epic.
[iv] Oracle Health Clinical AI Agent, Oracle.
[v] Acceptable Use of Artificial Intelligence Technologies, New York State Office of Information Technology Services, Updated 3/11/2025.
[vi] Summary of the HIPAA Privacy Rule, U.S. Department of Health and Human Services.
[vii] Is ChatGPT HIPAA Compliant?, The HIPAA Journal, April 9, 2025.
[viii] Building and Implementing an Artificial Intelligence Action Plan for Health Care, American Hospital Association, January 14, 2025.
[ix] Ibid.
[x] The Role of AI in Hospitals and Clinics: Transforming Healthcare in the 21st Century, Bioengineering (Basel), March 29, 2024.
[xi] SHIN-NY SCPA, NYeC.
[xii] The Role of AI in Hospitals and Clinics: Transforming Healthcare in the 21st Century, Bioengineering (Basel), March 29, 2024.
[xiii] AI at CommonSpirit: 230 Tools, $100M Impact, Becker’s Health IT, Sept. 2, 2025.
[xiv] Gross Financially Distressed Hospital Historic Support (FY 2017 – FY 2026), Health Care | Briefing Book | NYS FY 2026 Executive Budget, pg. 4.
[xv] Acceptable Use of Artificial Intelligence Technologies, NYS Office of Information Technology Services, Updated March 11, 2025.
[xvi] AI Transparency in the Age of LLMs: A Human-Centered Research Roadmap, Harvard Data Science Review Special Issue 5, February 29, 2024.
[xvii] Ethics and Governance of Artificial Intelligence for Health: Guidance on Large Multi-Modal Models, World Health Organization, 2024.
[xviii] The Role of AI in Hospitals and Clinics: Transforming Healthcare in the 21st Century, Bioengineering (Basel), March 29, 2024.
[xix] Ethics and Governance of Artificial Intelligence for Health: Guidance on Large Multi-Modal Models. World Health Organization, 2024.
[xx] Cost estimates were developed in private discussions with parties who are familiar with building this technology.
[xxi] Health Care | Briefing Book | NYS FY 2026 Executive Budget, p. 70.
[xxii] Health systems report multimillion-dollar returns from AI, Becker’s Health IT, October 1, 2025.