Blind Spots in Healthcare AI: Where Board Oversight Breaks Down

Updated October 21, 2026. First published August 16, 2023.

Boards rarely miss AI risk for lack of interest. They miss it because the risk sits where the board pack does not reach: vendor contracts, proxy variables, state statutes, clinicians who were never asked, and the months after go-live.

In Brief

The Problem. Healthcare, medical device, and life sciences companies built on artificial intelligence (AI) fail less often because the technology does not work than because leaders miss what surrounds it: bias, regulation, vendors, clinical adoption, patients, and proof of value.

Why It Persists. Board reporting describes the projects management brings forward. It rarely shows embedded AI, the proxy a model uses, the vendor's contract terms, or performance after launch, and these are where failures begin.

What Boards Should Do. Ask for an inventory that includes embedded AI, a named accountable executive, the rules that apply by jurisdiction and date, the vendor terms, evidence of clinician and patient involvement, measured value, and monitoring after go-live, and make sure the minutes show it.

In August 2023, I wrote about the blind spots that cause innovations in healthcare, AI, medical devices, and life sciences to fail. I addressed that article to boards and chief executives together, and most of its advice told management what to build. Three years later, the same blind spots have become questions of board oversight. AI now sits inside clinical documentation, claims, scheduling, imaging, and drug development. The rules governing it have multiplied, and the courts have clarified what directors and officers owe.

My argument is that most AI failures in healthcare are visible in advance to a board that knows where to look, and that the places to look are rarely the ones the technology discussion covers. This update keeps the original's blind spots, corrects them, sets them at board level, and adds the evidence that now exists for each. The posture is the one Avancer recommends throughout its work: nose in, fingers out. Directors do not need to evaluate models. They need to know what the organization uses, who is accountable, which rules apply, and what the minutes will show.

Understanding the Causes of Failure

The development of AI-enabled devices, clinical decision support tools, and new life sciences products sits at the frontier of medicine, and the field is crowded. The Food and Drug Administration (FDA) had authorized more than 1,400 AI-enabled devices as of its March 4, 2026 database update (Reuter & Han, 2026). In that crowd, leaders often focus narrowly on the technology and neglect the interplay between clinical utility, unmet need relative to the standard of care, and the realities of payer reimbursement and formulary decisions.

Three further causes recur in my experience. Leaders underestimate the time and resources needed for FDA clearance or approval abroad, which creates delays and financial strain. They treat ethical questions about AI, data privacy, and long-term social effects as matters to handle later, when a failure to anticipate them can quickly become reputational damage and lost trust. And they build products that are technically impressive but lack a clear value proposition, which, combined with weak go-to-market plans or gaps in specialized talent, undermines commercial viability. Each of these is now a governance question as well as a management one.

Why This Is a Board Matter

Under Delaware law, directors must make a good-faith effort to ensure a reporting system exists for the risks that matter most, and then monitor it. That duty traces to In re Caremark International Inc. Derivative Litigation (1996). In Marchand v. Barnhill (2019), the Delaware Supreme Court let an oversight claim proceed because the board had no system to monitor food safety, the company's central compliance risk (Marchand v. Barnhill, 2019). The Court of Chancery extended a comparable duty to officers, within each officer's area of responsibility, in In re McDonald's Corp. Stockholder Derivative Litigation (2023) (Youngwood et al., 2023).

The principle now reaches life sciences directly. In September 2025, the Court of Chancery let oversight claims proceed against directors of the drug maker Teligent, where the complaint alleged that no board committee was responsible for FDA compliance and that management decided whether and when to report compliance matters to the board (Bonadurer, 2026). For a hospital system, a payer, or a drug developer, AI that informs clinical, coverage, or regulatory decisions fits the description of a central risk. The standard is procedural. The board's best evidence of good faith is its own record.

Blind Spot One: The AI the Board Does Not Know About

Most boards can name the AI projects management has brought forward. Few can name the AI that arrived inside vendor products, such as ambient documentation in the electronic health record (EHR), denial prediction in revenue cycle software, and triage logic in the patient portal. These tools change when the vendor updates them, and the organization remains accountable for their effects. A vendor's performance claims also do not transfer automatically. When Michigan Medicine tested a proprietary sepsis prediction model implemented at hundreds of U.S. hospitals, it found an area under the curve of 0.63, against the 0.76 to 0.83 the developer reported; the model missed 67 percent of patients with sepsis while generating alerts in 18 percent of all hospitalizations (Wong et al., 2021).

The remedy is an inventory that includes embedded AI. A named executive should keep it current and report on it to a designated committee. Without an inventory, every other oversight question in this article rests on an incomplete list.

Blind Spot Two: Bias That Enters Through a Proxy

The 2023 article warned that algorithms can carry forward existing bias if they are not carefully trained and monitored. The best-known case shows how that happens without intent. Obermeyer and colleagues studied a widely used algorithm that allocated extra care management by predicting health costs (Obermeyer et al., 2019). Less money had historically been spent on Black patients with the same level of need, so the algorithm judged them healthier than equally sick White patients, and the bias cut the number of Black patients identified for extra care by more than half. Race was never an input; cost was the proxy.

This is now a legal question as well as an ethical one. In 2024, the Department of Health and Human Services (HHS) issued a rule under Section 1557 of the Affordable Care Act, at 45 CFR 92.210, requiring covered entities to make reasonable efforts to identify and mitigate discrimination risks from patient care decision support tools, with a compliance date of May 1, 2025; the duty cannot be delegated to the tool's developer (Han et al., 2024). Practitioners noted in April 2025 that enforcement under the current administration is uncertain (Pressly & Leeds, 2025), and boards should confirm the rule's current status with counsel. States are moving regardless. California's SB 503, effective January 1, 2027, requires developers and deployers of clinical decision support systems to make reasonable efforts to identify tools that could produce biased results, and requires deployers to keep monitoring their effects (Montano & Tran, 2026).

The original article recommended testing across diverse populations and multi-stakeholder review boards that include clinicians, ethicists, data scientists, patient representatives, and community advocates. Both remain sound, and the Joint Commission and the Coalition for Health AI (CHAI) now list governance by a multidisciplinary body and risk and bias assessment on representative data among the elements of responsible use (Polsinelli, 2025). The board's questions are simple. What does each clinical tool predict? What does it use as a stand-in for need? How did the organization test its results across patient groups, and who reviewed them?

Blind Spot Three: The Rulebook Changing Under the Deployment

The 2023 article advised treating compliance as part of design, not an afterthought, and engaging regulators early. That advice holds, and an experienced regulatory affairs function remains essential. For AI-enabled devices, the FDA finalized its guidance on predetermined change control plans in December 2024, allowing manufacturers to make preapproved modifications without a new authorization for each significant change (Organ et al., 2025). The FDA's January 2025 draft guidance on lifecycle management of AI-enabled device software functions remains in draft (U.S. Food and Drug Administration, 2025).

The larger change is at the state level. In May 2026, Colorado repealed and replaced its 2024 AI Act. The new law takes effect on January 1, 2027, exempts entities covered by the Health Insurance Portability and Accountability Act (HIPAA) from many of its obligations, and still requires a general patient notice about the use of advanced technologies (Darch et al., 2026). Federal preemption remains unresolved. A December 11, 2025 executive order told the Department of Justice (DOJ) to form a task force to challenge state AI laws (Paul Hastings LLP, 2025), and on April 24, 2026, the DOJ moved to join a suit by xAI challenging the earlier Colorado law (Cooper et al., 2026). State laws remain in force unless a court rules otherwise.

Companies that sell abroad face a third calendar. The European Union's AI Omnibus, Regulation (EU) 2026/1744, entered into force on July 27, 2026, and deferred high-risk obligations for standalone systems to December 2, 2027 and for AI embedded in regulated products, such as medical devices, to August 2, 2028 (Pearce & Nietsch, 2026). The board's question is practical: for each deployment or product, which rules apply where it operates, from what date, and who in management tracks the changes?

Blind Spot Four: The Vendor Who Holds the Risk

The 2023 article treated cybersecurity as a technical competency: employee training on phishing and social engineering, regular penetration testing and vulnerability assessments, detailed incident response plans, and transparent breach communication with staff and patients. Those practices remain necessary. The February 2024 attack on Change Healthcare showed that cybersecurity is also a concentration risk; on July 31, 2025, Change Healthcare notified the Office for Civil Rights (OCR) that approximately 192.7 million individuals were affected (U.S. Department of Health and Human Services, Office for Civil Rights, 2025). The federal backstop is moving slowly: OCR has delayed its final HIPAA Security Rule, proposed in January 2025, to a July 2027 target (Everett et al., 2026).

AI deepens the dependence. Each deployment adds a vendor, a data flow, and often a model the organization cannot inspect. In September 2024, the Texas attorney general settled with Pieces Technologies, whose generative AI product summarized patients' conditions and treatments for staff in at least four major Texas hospitals. The investigation found that the company's advertised severe hallucination rate of less than one per 100,000 was likely inaccurate, and the settlement requires the company to disclose its products' accuracy and to ensure that hospital staff understand how far to rely on them (Office of the Texas Attorney General, 2024). Integration is part of the same risk. The original article was right that isolated tools fail without integration into EHR platforms and data exchange standards such as Health Level Seven (HL7) Fast Healthcare Interoperability Resources (FHIR). But the transparency boards might have expected from certified health IT is receding: the HTI-5 proposed rule of December 29, 2025 would remove the requirement that developers supply "model card" information and certain risk management practices for predictive decision support tools (Covington & Burling LLP, 2026). What regulation does not require, the contract must.

AI approvals should therefore come to the board with the vendor terms that matter. May the vendor train on the organization's data? How will it give notice of model changes? What audit rights does the organization hold, and how is indemnification allocated? Directors should also ask which single vendors, if lost, would interrupt clinical or financial operations.

Blind Spot Five: The Clinicians Who Were Not Ready

Even the most promising technology will falter if the clinicians who use it are not equipped and were not consulted. Use has outrun preparation. In a March 2026 American Medical Association (AMA) survey, 81 percent of physicians reported using AI in their practices, up from 38 percent in 2023; 88 percent were concerned about potential skill loss, 88 percent said safety and efficacy validation is critical, and 85 percent wanted to be consulted or involved in adoption decisions (American Medical Association, 2026). California's AB 1979, effective January 1, 2027, requires facilities and physician offices to take reasonable steps to ensure a licensed provider exercises independent judgment whenever AI output informs care (Montano & Tran, 2026).

The original article's remedies still apply: training that covers judgment as well as technical use, change management that shows how a tool fits existing workflows, continuing education with universities, and in-house "train-the-trainer" models. The board does not design these programs. It should ask whether clinicians helped select the tool, how they are trained to question its output, and how often they override it.

Blind Spot Six: The Patient Who Gets Lost

Healthcare innovation exists to serve patients. Yet too often, the patient's needs are lost in the pursuit of technological advancement. The 2023 article urged leaders to treat patients as partners: patient advisory councils with input at every stage of development, feedback on prototypes, usability, trial design, and instructions for use, qualitative work on the lived experience of patients and caregivers, and compensation for patients' time and expertise. Those practices stand, and they now carry a disclosure dimension. The Joint Commission and CHAI guidance calls for telling patients when AI is used and obtaining informed consent when AI meaningfully influences treatment decisions (Polsinelli, 2025).

Access is the other half of this blind spot. New technologies can widen health disparities if access is limited to the well served. Digital tools assume digital readiness, and many patients lack it: an analysis of national survey data from 2018 estimated that 38 percent of Americans aged 65 and older, about 13 million people, were not ready for video visits, mainly because of inexperience with technology rather than disability, rising to about 72 percent of those aged 85 and older (Lam et al., 2020). The original remedies remain relevant: tiered pricing; partnerships with public and nonprofit payers serving underserved communities; digital health literacy programs with libraries, senior centers, and local clinics; translated materials and accessibility features; and open-source elements in core technologies where they help local adaptation. The board's question is who is left out by each deployment, and what the plan is for them.

Blind Spot Seven: Value Claimed, Not Shown

The 2023 article noted that healthcare is moving from fee-for-service toward payment tied to outcomes. The shift is substantial but incomplete. In calendar year 2024, 44.9 percent of U.S. health care payments flowed through alternative payment models, and 28.7 percent through models with downside risk (AHIP, 2026). In that market, a product must show value beyond clinical endpoints: quality of life (pain, function, mental health), caregiver burden, readmissions, time away from work, and cost offsets for the health system, ideally measured with validated patient-reported outcome measures (PROMs) and metrics co-designed with payers and providers.

The same discipline applies to AI the organization buys. In a September 2026 survey of 226 healthcare executives, buyers reported an average return of 3.5 times on AI spending to date and payback in about 12 months, led by administrative uses such as revenue cycle, but the figures are self-reported (S. Guerra & Kraus, 2026). Boards should ask for value measured by use case against a baseline set before deployment, not adoption counts.

Blind Spot Eight: The Months After Go-Live

The 2023 article urged teams to collect feedback from clinicians, patients, and real-world evidence (RWE) to refine products, and to plan regulatory submissions for modular updates. For a board, the blind spot is that approval is treated as the end of oversight. I have seen it happen. It is the beginning. Models drift as patient populations, documentation habits, and vendor versions change, and clinicians who were not trained to question an output will tend to accept it.

The Joint Commission now offers an external benchmark. Its voluntary Responsible Use of AI in Healthcare certification launched on June 1, 2026, and assesses governance, data management, risk and bias reduction, monitoring and validation of safety performance, and transparency, education, and training (A. Guerra, 2026). Whether or not the organization pursues certification, those areas describe what management should report on a schedule: performance against the validation baseline, incidents, overrides, and any decision to pause or retire a tool.

The Wider Footprint: Supply Chains, Sustainability, Intellectual Property, and Global Reach

Five of the original article's blind spots sit outside AI itself but still belong on the board's agenda, because AI products depend on them.

Technology foresight. The pace of technological change can outstrip a company's ability to adapt. The 2023 article recommended a dedicated horizon-scanning function and partnerships with academic laboratories, incubators, and early-stage companies to find applications that could disrupt or complement existing products. That remains management's work. The board's question is who tells it what is coming, how often, and how those findings reach the strategy discussion.

Supply chain resilience. The vulnerabilities exposed by the COVID-19 pandemic have not gone away. The American Society of Health-System Pharmacists (ASHP) counted 227 active drug shortages in the second quarter of 2026, after a peak of 323 in the first quarter of 2024 (Shaw, 2026). The original advice holds: avoid sole suppliers for critical components, diversify by geography, and stress-test the supply chain as part of scenario planning that also covers pandemics, large data breaches, and natural disasters. The board should see the results of those exercises, not only know they occurred.

Sustainability. The healthcare sector contributes materially to pollution; U.S. health care accounted for about 8.5 percent of domestic greenhouse gas emissions in 2018 (Eckelman et al., 2020). Designing devices for their full lifecycle, with sustainable materials, repairability, safe disposal, and measured energy use, waste, and carbon footprint, remains sound practice and a reasonable subject for periodic board reporting.

Intellectual property. Intellectual property is a core asset in these industries, and AI has complicated it. In Thaler v. Vidal (2022), the Federal Circuit held that only natural persons can be named as inventors under the Patent Act. Companies that use AI in discovery and design therefore need clear protocols for documenting human contribution, publications, and collaborations, and regular audits of their patent, trademark, and trade secret positions.

Global reach and collaboration. The 2023 article urged collaboration across disciplines, sectors, and countries, including co-development with partners in low- and middle-income countries and in-house teams that bring clinicians, engineers, data scientists, designers, and patient advocates together. That remains a strategic choice for management. For the board, the governance consequence is the jurisdictional map described under the third blind spot: every new market adds a rulebook and a calendar.

The Strongest Objection: This Is Management's Work

The strongest objection is that most of this list is management's job. Directors are not equipped to judge model validation, interoperability standards, or supply chain software, and a board that tries will slow the organization and blur accountability.

The objection is right about execution and wrong about oversight. Nothing here asks directors to do the work. It asks them to know that the work exists, who owns it, and how they will hear when it fails. That is precisely the line the Delaware cases draw: liability attaches to the absence of a reporting system for central risks, not to a board's failure to run them (Marchand v. Barnhill, 2019). The record will also be read by others. In a pending class action alleging that UnitedHealth used an AI tool to deny Medicare Advantage coverage, a claim the company disputes, a federal magistrate judge in March 2026 ordered discovery that includes materials about the company's internal AI review board (Greenberg et al., 2026). Governance documents are evidence.

The Leadership Mandate

Leaders of healthcare innovation companies shape the future of patient care, the trajectory of health equity, and what responsible technological progress means. Every strategic decision and every resource allocation echoes far beyond the boardroom. Regulatory missteps delay treatments for patients who need them, and ethical lapses in AI and data use erode public trust in innovation for the entire industry.

Each blind spot leaves a trace in the record once the board addresses it. The minutes should show an inventory reviewed, an accountable executive named, applicable rules listed by jurisdiction and date, vendor terms summarized, clinician and patient involvement described, value measured against a baseline, conditions attached to each approval, and monitoring reports received. That record is what a court, a regulator, or an accreditor will read.

For the Board

• Does our AI inventory include AI embedded in vendor products, and who is accountable for keeping it current?
• For each clinical or coverage tool, what does it predict, what proxy does it use, and how did we test it across patient groups?
• Which state, federal, and foreign rules apply to each deployment or product, from what dates, and has counsel confirmed the current status of the Section 1557 tool rule?
• Do our vendor contracts address training on our data, notice of model changes, audit rights, and indemnification, and which single vendors could interrupt operations?
• Were clinicians and patients involved in selecting and designing the tool, and who is left out by it?
• What value and performance will management report to us after go-live, against what baseline, and what would cause us to switch a tool off?

Board question sets for each of these blind spots, organized by committee, are in the Avancer AI Governance Navigator.

References

AHIP. (2026). 2025 measurement effort [Interactive infographic]. Retrieved October 21, 2026, from https://view.ceros.com/ahip/2026-01-interactive-infographic-1-1

American Medical Association. (2026, March 12). AMA: AI usage among doctors doubles as confidence in technology grows [Press release]. https://ama-assn.org/press-center/ama-press-releases/ama-ai-usage-among-doctors-doubles-confidence-technology-grows

Bonadurer, J. (2026, February). Delaware court denies dismissal of Caremark oversight claims in Teligent case, highlighting board and officer compliance obligations. Jenner & Block LLP. https://www.jenner.com/en/news-insights/publications/delaware-court-denies-dismissal-of-caremark-oversight-claims-in-teligent-case-highlighting-board-and-officer-compliance-obligations

Cooper, A. R., Weaver, W., Cease, C., & Ramsey, A. R. P. (2026, April 29). DOJ joins xAI in lawsuit challenging Colorado AI Act. Jenner & Block LLP. https://www.jenner.com/en/news-insights/client-alerts/doj-joins-xai-in-lawsuit-challenging-colorado-ai-act

Covington & Burling LLP. (2026, January 15). HHS proposes changes to the health IT certification program and information blocking regulations in HTI-5 proposed rule. https://www.covingtonblogs.com/2026/01/15/hhs-proposes-changes-to-the-health-it-certification-program-and-information-blocking-regulations-in-hti-5-proposed-rule/

Darch, J. E., Brooks, P., Daudi, S., Hinckley, D., & Kaye, A. (2026, May 19). Colorado scales back AI law with targeted implications for health care. Ropes & Gray LLP. https://www.ropesgray.com/en/insights/alerts/2026/05/colorado-scales-back-ai-law-with-targeted-implications-for-health-care

Eckelman, M. J., Huang, K., Lagasse, R., Senay, E., Dubrow, R., & Sherman, J. D. (2020). Health care pollution and public health damage in the United States: An update. Health Affairs, 39(12), 2071 to 2079. https://doi.org/10.1377/hlthaff.2020.01247

Everett, J. C., Burnette, A. T., & Pike, J. (2026, July 10). HHS Office for Civil Rights delays final HIPAA Security Rule until 2027 as Privacy Rule changes near. Alston & Bird LLP. https://www.alston.com/en/insights/publications/2026/07/hhs-delays-hipaa-security-rule-2027

Greenberg, D. S., Heilman, K., & Shlaferman, H. Z. (2026, April 23). Federal court orders broad discovery against UHC in AI coverage denial lawsuit. ArentFox Schiff LLP. https://www.afslaw.com/perspectives/alerts/federal-court-orders-broad-discovery-against-uhc-ai-coverage-denial-lawsuit

Guerra, A. (2026, June 1). Joint Commission launches AI certification as adoption outpaces oversight. HealthSystemCIO. https://healthsystemcio.com/2026/06/01/joint-commission-ai-certification/

Guerra, S., & Kraus, S. (2026, September 30). The 2026 healthcare AI ROI scorecard. Bessemer Venture Partners. https://www.bvp.com/atlas/the-2026-healthcare-ai-roi-scorecard

Han, J., Tsai, J., & Khazanchi, R. (2024, November 14). Section 1557 rule mandates identification and mitigation of discriminatory clinical algorithms. Health Affairs Forefront. https://doi.org/10.1377/forefront.20241113.513530

In re Caremark International Inc. Derivative Litigation, 698 A.2d 959 (Del. Ch. 1996).

In re McDonald's Corp. Stockholder Derivative Litigation, 289 A.3d 343 (Del. Ch. 2023).

Lam, K., Lu, A. D., Shi, Y., & Covinsky, K. E. (2020). Assessing telemedicine unreadiness among older adults in the United States during the COVID-19 pandemic. JAMA Internal Medicine, 180(10), 1389 to 1391. https://doi.org/10.1001/jamainternmed.2020.2671

Marchand v. Barnhill, 212 A.3d 805 (Del. 2019). https://law.justia.com/cases/delaware/supreme-court/2019/533-2018.html

Montano, M., & Tran, S. (2026, October 8). Newsom approves 3 health care AI bills, vetoes 2. California Health Care Foundation. https://chcf.org/blog/2026/10/08/newsom-approves-3-health-care-ai-bills-vetoes-2

Obermeyer, Z., Powers, B., Vogeli, C., & Mullainathan, S. (2019). Dissecting racial bias in an algorithm used to manage the health of populations. Science, 366(6464), 447 to 453. https://doi.org/10.1126/science.aax2342

Office of the Texas Attorney General. (2024, September 18). Attorney General Ken Paxton reaches settlement in first-of-its-kind healthcare generative AI investigation [Press release]. https://www.texasattorneygeneral.gov/news/releases/attorney-general-ken-paxton-reaches-settlement-first-its-kind-healthcare-generative-ai-investigation

Organ, S. A., Beaver, N. A., & Rosen, D. L. (2025, January 6). AI-enabled device software functions: FDA's final guidance for predetermined change control plans. Foley & Lardner LLP. https://www.foley.com/insights/publications/2025/01/fda-final-guidance-ai-device-software-predetermined-change-control-plan/

Paul Hastings LLP. (2025, December 16). President Trump signs executive order challenging state AI laws. https://www.paulhastings.com/insights/client-alerts/president-trump-signs-executive-order-challenging-state-ai-laws

Pearce, S., & Nietsch, T. (2026, July 31). EU Digital Omnibus on AI enters into force. The National Law Review. https://natlawreview.com/article/eu-digital-omnibus-ai-enters-force

Polsinelli. (2025, October 29). Joint Commission and Coalition for Health AI release first-of-its-kind guidance on responsible AI use in healthcare. https://www.polsinelli.com/publications/tjc-chai-release-guidance-responsible-ai-use-healthcare

Pressly, D. F., & Leeds, E. I. (2025, April 17). Section 1557 deadlines approach, but long-term prognosis unclear under executive orders. Ballard Spahr LLP. https://www.ballardspahr.com/insights/alerts-and-articles/2025/04/section-1557-deadlines-approach-but-long-term-prognosis-unclear-under-executive-orders

Reuter, E., & Han, J. Y. (2026, May 27). AI in medtech is booming. Track new devices here. MedTech Dive. Retrieved October 21, 2026, from https://www.medtechdive.com/news/ai-medtech-track-new-devices-fda/748397/

Shaw, M. L. (2026, July 20). Active US drug shortages rise for third straight quarter. The American Journal of Managed Care. https://www.ajmc.com/view/active-us-drug-shortages-rise-for-third-straight-quarter

Thaler v. Vidal, 43 F.4th 1207 (Fed. Cir. 2022). https://www.bitlaw.com/source/cases/patent/Thaler.html

U.S. Department of Health and Human Services, Office for Civil Rights. (2025). Change Healthcare cybersecurity incident frequently asked questions. Retrieved October 21, 2026, from https://www.hhs.gov/hipaa/for-professionals/special-topics/change-healthcare-cybersecurity-incident-frequently-asked-questions/index.html

U.S. Food and Drug Administration. (2025, January). Artificial intelligence-enabled device software functions: Lifecycle management and marketing submission recommendations (Draft guidance). Retrieved October 21, 2026, from https://www.fda.gov/regulatory-information/search-fda-guidance-documents/artificial-intelligence-enabled-device-software-functions-lifecycle-management-and-marketing

Wong, A., Otles, E., Donnelly, J. P., Krumm, A., McCullough, J., DeTroyer-Cooley, O., Pestrue, J., Phillips, M., Konye, J., Penoza, C., Ghous, M., & Singh, K. (2021). External validation of a widely implemented proprietary sepsis prediction model in hospitalized patients. JAMA Internal Medicine, 181(8). https://doi.org/10.1001/jamainternmed.2021.2626

Youngwood, J. K., Goldin, N., & Blake, S. (2023, February 7). Corporate officers, not just directors, can be liable for duty of oversight violations. Harvard Law School Forum on Corporate Governance. https://corpgov.law.harvard.edu/2023/02/07/corporate-officers-not-just-directors-can-be-liable-for-duty-of-oversight-violations/