When Technology Becomes Obligation: AI and Duty

Saint Thomas Aquinas taught that a person may be responsible for neglecting knowledge needed to fulfill a duty.¹ The teaching belongs to another century, yet it carries a modern resonance. We live in a moment when knowledge arrives before illness announces itself, when patterns inside the body are revealed long before a human eye could notice them. Artificial intelligence has entered healthcare. It is being applied in radiology and pathology, studied in cardiology and dermatology, and extended into ordinary life through predictive analytics and continuous wearables.² These tools offer earlier detection and a kind of ever-present watchfulness that previous generations of medicine could not imagine.

Their rise invites a renewed question about knowledge and responsibility. When reason is amplified by technology, and when that technology uncovers danger in its earliest traces, at what point does the ability to know become the responsibility to act? When does a tool that sharpens our sight turn into a duty that law can recognize?

This paper argues that artificial intelligence in early diagnosis and preventive medicine will expand the demands of reasonable care for healthcare providers and reshape the expectations placed upon patients. Decisions about diagnostic precautions already show how a new means of preventing harm can change what reasonable care requires.³

The idea that advancing technology can reshape legal responsibility is far older than artificial intelligence. Blackstone described injury through a physician's neglect or unskillful treatment as a breach of the trust a patient had placed in the physician.⁴ Duty has never been a static command. It responds to the practical abilities of ordinary life. When the means of avoiding harm become clearer or easier, conduct once accepted as unavoidable comes into question.

This concern appears even in the Case of Thorns from 1466. While the defendant was cutting a hedge, thorns fell onto his neighbor's land; he entered to retrieve them. He argued that the thorns had fallen against his will. Choke J. reasoned that this was not enough: the defendant had to show that he could not have acted otherwise or had done all he could to prevent the fall.⁵ A lawful purpose and an unintended injury did not end the inquiry. The question was what the defendant could have done to avoid the harm.

Judge Buller later expressed the rule in more precise form in his eighteenth-century writings, stating that “every man ought to take reasonable care that he does not injure his neighbour.”⁶ These early thoughts express a principle with continuing force: as people gain better means of prevention, the law rightly expects greater caution. Buller's framing expresses a foundation of negligence doctrine: a duty to exercise reasonable care with the knowledge and methods already within reach.

By the nineteenth century, courts were openly describing professional responsibility in terms of the ordinary learning, skill, and experience of the profession. In Leighton v. Sargent, the court explained that a surgeon undertook to possess those qualifications and to apply them with reasonable care and diligence.⁷ A standard measured by professional knowledge cannot be frozen while that knowledge advances. As the instruments of diagnosis and treatment improve, the practical content of competent care can advance with them. This historical pattern shows that reasonable care is not a rigid formula. Rather, it is an evolving measure that grows with human capability. The more we can foresee, the greater the responsibility to prevent. The more precise our tools, the sharper the line between acceptable risk and preventable harm. It is a change in the field of vision. It transforms what people can know, and therefore what they ought to avoid.

The pattern becomes especially clear when a precaution is inexpensive, low-risk, and effective against serious harm. In Helling v. Carey, the Washington Supreme Court required a glaucoma test even though professional custom did not call for it for patients of that age. Harris v. Groth later explained that professional practice is evidence of reasonable care rather than its conclusive measure.⁸ The practical lesson reaches beyond any single instrument: the habits of a profession must be considered alongside the harm its available tools can prevent.

Artificial intelligence enters this lineage naturally. Its power to detect patterns, predict risk, and reveal emerging danger is the newest expansion in human capability. The law has adapted to changes in what people can reasonably know and do, and it is being asked to do so again in the age of digital diagnosis. AI is the next chapter in a very old story of how law responds to the growth of human understanding. As knowledge arrives through the quiet grace that accompanies discovery, the excuses for preventable harm begin to diminish.

The evolution of neonatal medicine offers one of the clearest illustrations of how expanding capability reshapes the boundaries of duty. In Bell and colleagues' study of nineteen United States centers from 2013 through 2018, survival among infants actively treated at birth reached 30.0 percent at twenty-two weeks, 55.8 percent at twenty-three weeks, and 71.4 percent at twenty-four weeks. The outcome was survival to hospital discharge, or to one year for infants still hospitalized.⁹ This progress is reflected in Guinness World Records' recognition of Nash Keen, born in 2024 at exactly twenty-one weeks' gestation, as the most premature infant to survive.¹⁰ His individual milestone shows how far neonatal medicine has pushed the boundary of possibility. A central barrier at these gestations remains the immaturity of the lungs. Their gas-exchange structures are still developing, and the oxygen and mechanical support needed to sustain life can themselves injure fragile tissue. Much of modern neonatology has been devoted to finding ways around this biological limit.¹¹

Improving survival has accompanied changes in the way premature infants are treated. In the earlier 1993–2012 cohort studied by Stoll and colleagues, antenatal corticosteroid use increased and less-invasive respiratory support became more common. Survival improved most markedly at twenty-three and twenty-four weeks.¹² Together, advances in care have brought children once beyond medicine's reach into the realm of survivability. Neonatal medicine is pursuing ways to give immature organs the time and conditions they need to mature.

The most striking experimental development is the creation of external-womb systems. Research teams have supported fetal lambs in an artificial amniotic environment, often called a “Biobag,” for as long as twenty-eight days. These systems mimic the fluid-filled conditions of the uterus and allow continued development without relying on ventilation through immature lungs.¹³ Parallel efforts to build an artificial placenta transfer oxygen through an extracorporeal circuit. In one study, six extremely preterm fetal lambs were supported for an average of 250 hours, with a maximum of 336 hours, although growth remained reduced.¹⁴

These experiments approach the lung barrier from a different direction. They preserve fetal circulation and fluid-filled lungs while development continues outside the uterus.¹⁵ The ambition reaches beyond helping immature lungs breathe: it is to sustain the conditions under which those lungs can mature. If such technologies mature into safe clinical practice and can be adapted to earlier gestational ages, they could shift viability downward to ages previously beyond medicine's reach.

As these capabilities grow, the meaning of viability itself begins to shift. When Roe v. Wade was decided, the Court described viability as usually occurring around twenty-eight weeks, while acknowledging that it could occur at twenty-four.¹⁶ Planned Parenthood v. Casey later acknowledged explicitly that advances in neonatal care had moved viability earlier.¹⁷ Debates over whether the decisive line rests at twenty-four weeks, twenty-two weeks, or somewhere earlier reflect not a change in metaphysics but a change in capability. Viability is not a fixed boundary. It is a measurement of human skill at a particular moment in history. A judgment that an unborn child cannot be saved is therefore partly a judgment about the knowledge and tools available at that moment. As neonatal medicine progresses, so does the horizon of what life can be sustained, and with it, the responsibilities tied to that possibility.

John Paul II described healthcare professionals as “guardians and servants of human life.”¹⁸ Neonatal technology illustrates this truth. Each expansion of capability brings a corresponding expansion of responsibility, a pattern mirrored in artificial intelligence's expansion of medical foresight. Does the same technological trajectory that reshapes viability reshape the expectations placed upon those who practice?

If the history of negligence law teaches anything, it is that reasonable care changes when the means of prevention change. Nowhere is that more evident than in medicine, where the standard of care is inseparable from the tools that help clinicians diagnose, foresee, and respond to danger. Artificial intelligence intensifies this connection. It widens the range of what can be known, and in doing so, it changes what a reasonable provider must consider.

Modern malpractice law already frames this expectation. Statutes such as section 766.102 of the Florida Statutes define the medical standard of care as the “level of care, skill, and treatment” “recognized as acceptable and appropriate by reasonably prudent” similar providers.¹⁹ What is reasonably available matters to that inquiry.²⁰ Consider the difference between a cardiologist who lacks access to an echocardiogram and one who ignores significant results placed directly into her hands. Technology informs what prudence requires.

Artificial intelligence magnifies this relationship because it transforms the raw material of medical knowledge. In one study, a deep neural network matched board-certified dermatologists on two skin-lesion classification tasks. In cardiology, a retrospective study used an AI-enabled electrocardiogram algorithm to identify signs of atrial fibrillation in recordings made during normal sinus rhythm. Documentation tools can structure clinical information and help detect errors in medical notes.²¹ Such capabilities change what a provider may be able to discover. Once a reliable tool becomes appropriate to a clinical setting, that additional knowledge can bear on what a court later says the provider should have known.

The division of responsibility matters as much as the instrument. In Sampson v. HeartWise Health Systems, HeartWise generated reports from screening data, while interpretation, diagnosis, and referral remained with the treating physicians. The Alabama Supreme Court affirmed summary judgment for HeartWise on the negligence claims and remanded those claims against Isaac Health and its physicians because Isaac Health and its physicians had never moved for summary judgment on those negligence claims.²² The arrangement illustrates why a technology's place in care must be understood through the work each participant undertakes. Generating a report and deciding what to do with it are different responsibilities. The tool may assist, but clinical judgment remains a human task.

As AI becomes more reliable, courts may begin to treat it the way they once treated diagnostic tests. Helling v. Carey offers a concrete example. A patient who had consulted her ophthalmologists over several years was diagnosed with glaucoma at thirty-two, after substantial visual damage. Although the profession did not ordinarily require tonometry testing for patients under forty, the pressure test was simple, inexpensive, and harmless when the patient's eye condition permitted its use, and early detection could permit treatment to arrest the disease. The court held that custom did not excuse the failure to give this patient the test.²³ The reasoning was unapologetically functional. The profession's habits could not override the demands of reasonable care.

It is not difficult to imagine a parallel in AI-supported diagnostics. If a validated artificial intelligence system consistently outperforms human interpretation in a specific setting, and if that system is accessible, affordable, and integrated into routine clinical practice, courts may view its omission much like the omission in Helling.²⁴ Custom alone will not settle the question. Capability will matter too. A tool that reliably illuminates danger can become a tool that reasonable care requires.

The story of Helling did not end with the decision. Washington enacted medical-negligence statutes, and Harris later held that reasonable prudence remained the governing standard, with professional practice serving as evidence of it. Florida draws a particular boundary for supplemental diagnostic tests: section 766.102(4) protects a provider who omits such tests while acting in good faith and with due regard for the prevailing professional standard of care.²⁵ These rules shape how the argument must operate in a particular jurisdiction. They also keep the central inquiry in view: what does reasonable medical care require when the means of prevention change?

The shift will not happen overnight. Human oversight and individualized judgment remain central to competent care. But the trajectory is familiar. If validated AI makes certain forms of early detection both accurate and readily available, the standard of care can rise accordingly. Providers will then be expected to use those tools with competence, to verify their outputs, and to integrate them into clinical reasoning. In time, failure to consult an appropriate system may be viewed much like failure to order an indicated test or attend to a clear warning.

Artificial intelligence does not displace provider duty. It expands the means of fulfilling it. It reveals more, earlier, and with fewer barriers. And as useful knowledge moves within reach, the law will increasingly expect clinicians to use the full measure of what it offers.

The expansion of medical responsibility has never belonged to providers alone. Long before artificial intelligence entered clinical practice, courts considered whether patients' own choices contributed to their injuries.²⁶ This responsibility reflects agency. A patient who understands a danger and receives reasonable instructions for avoiding it participates in the work of prevention. Artificial intelligence widens what patients can know and how easily they can know it. That makes the patient's role in care increasingly important.

Comparative negligence allows courts to allocate fault between provider and patient when both contribute to an injury.²⁷ In medical cases, a patient's unreasonable failure to follow medical advice, return for recommended care, or disclose relevant symptoms can affect recovery when it contributes to the harm.²⁸ The law does not expect perfection. It expects reasonable attention to one's own safety.

Dunnington v. Virginia Mason illustrates the significance of those choices. The patient initially chose cryotherapy over surgical excision for a persistent lesion. At follow-up, the physician discussed excision and biopsy, favored excision, and instructed him to return. He again chose conservative treatment and did not return as instructed. After a later MRI, he deferred an excisional biopsy, sought another opinion, and again chose conservative care. A subsequent biopsy revealed melanoma; the cancer recurred after treatment, and part of his leg was ultimately amputated. The Washington Supreme Court held that disputed facts required the contributory-negligence defense to go to the factfinder.²⁹ Choice carries consequence when it materially contributes to preventable harm.

Witherell v. Weimer presented another example. The defense alleged that the patient refilled prescriptions without approval, ignored dosage instructions, failed to reduce medication when advised, took long automobile trips contrary to medical warnings, and delayed reporting leg pain. A jury attributed forty percent of her injury to her own conduct. The Illinois Supreme Court found support for at least some of the asserted grounds and no reversible error in that allocation.³⁰ The case stands for a straightforward proposition. Compliance with reasonable instructions is part of the legal fabric of medical care.

These precedents do not treat the patient as a passive subject. They treat the patient as a participant whose choices matter. A reasonable patient must respond when a physician identifies danger, recommends accessible care, or offers clear instructions for follow-up. That responsibility supplies a reason to expect engagement with tools within easy reach, especially when those tools are inexpensive, available, and capable of preventing serious harm. On this account, responsibility is reciprocal because the relationship is reciprocal.

Artificial intelligence enlarges this field. Researchers have used a commercially available smartwatch and a deep neural network to identify atrial fibrillation. Retrospective analysis of smartwatch data has also detected physiological changes preceding reported COVID-19 symptoms in some participants.³¹ A watch worn in ordinary life can thus become a source of information about changes otherwise unnoticed. When that information is developed into a reliable, understandable warning connected to medical advice, it can change what a patient reasonably should do.

The legal implications follow naturally from the historical pattern. If a patient receives a clear, clinically reliable alert that signals risk, ignoring it may resemble ignoring a doctor's direct instruction. If a patient refuses to use a simple monitoring device that a provider reasonably recommends, and that decision contributes to a preventable harm, courts may consider the refusal in allocating responsibility. The analogy to the seatbelt defense is instructive. A seatbelt does not prevent every injury, but it can reduce harm when an accident occurs. In Spier v. Barker, New York permitted nonuse of an available seatbelt to bear on damages when the defendant proved that it increased the injuries.³² AI-enabled monitoring may in time stand in a similar place: an accessible precaution whose unreasonable neglect worsens an injury. It cannot prevent every illness, but it may reveal some risks sooner and with little effort.

Artificial intelligence gives new force to a familiar idea. Responsibility grows as understanding grows. When technology reveals danger earlier and more clearly, the patient's reasons to respond become stronger. Reasonable care does not demand unrealistic vigilance. It accounts for access to the tool and the ability to understand the warning. It should expect patients to engage with useful medical guidance and take simple steps that meaningfully reduce harm. Duty expands with capability. For patients, artificial intelligence increases capability in ways the law cannot ignore.

Public policy is helping to shape the conditions in which artificial intelligence enters medical care. In 2025, federal policy promoted AI development while the Food and Drug Administration proposed guidance for evaluating the safety and effectiveness of AI-enabled medical devices.³³ Florida offers a concrete example of the research infrastructure developing alongside that policy. The OneFlorida+ Clinical Research Network brings together academic institutions and health systems to support large-scale research using health data.³⁴ Such infrastructure helps build the knowledge from which future clinical practices can grow.

One of Florida's most visible commitments is the University of Florida's partnership with NVIDIA. Announced in 2020, the $70 million initiative combined a $50 million gift with a further $20 million investment from UF. It expanded HiPerGator's computing capacity and linked AI research with curriculum development and healthcare applications. UF researchers also developed DeepSOFA, a model using longitudinal electronic-health-record data to predict in-hospital mortality. It was tested across two hospital populations and externally validated.³⁵ These projects bring predictive modeling, professional education, and patient-care research into the same institutional setting.

The significance of Florida's AI infrastructure is not merely technical. It is doctrinal. When a tool becomes a reliable part of clinical education and practice, it can become part of what courts examine as ordinary professional competence. A training environment that teaches future clinicians to evaluate and use predictive tools begins to define the contours of reasonable practice. As research becomes effective practice, the argument that a prudent clinician should understand and respond to these systems becomes stronger.

Florida's telehealth framework offers a related principle. It requires telehealth providers to practice consistently with the prevailing professional standard applicable to in-person care.³⁶ This principle of parity has broad implications. It supports the inference that delivering care through technology does not diminish professional responsibility. As AI becomes part of clinical work, the same concern will arise: how should a provider use the available means to deliver competent care?

By supporting these technologies through research, education, and clinical infrastructure, public institutions help construct a world in which patient participation can become a natural part of technology-assisted care. The law of duty evolves in the landscape that the state itself helps to build.

Florida's combination of computing resources, medical research, and professional education gives it the potential to shape future expectations for providers and patients. As useful AI tools become established in particular clinical settings, their presence can inform what reasonable care requires. The regulatory and educational choices made today will help shape the practices courts examine tomorrow. By building the conditions in which technology becomes part of ordinary care, the state helps determine when technology becomes obligation.

Aquinas taught that the knowledge our duties require cannot be carelessly ignored, and the history of negligence law reflects a related idea. This paper has traced how reasonable care responds when new tools make prevention more attainable, from early common-law reasoning to modern diagnostic precautions. Artificial intelligence now stands in that same lineage. For providers, reliable diagnostic and predictive tools enlarge what can be known about a patient's condition and therefore what reasonable care can demand. For patients, accessible monitoring and clear medical warnings give new force to the expectation that they participate in their own safety. Florida's research and educational investments show how institutions can help shape those expectations. AI may extend traditional principles by enlarging human capability. As the means to foresee and prevent harm grow, so may the duty to use them, and the law will move in that direction for physicians and patients alike.

Alan Turing, the mathematician whose work helped define the contours of artificial intelligence, observed that we can only see a short distance ahead, yet we can see plenty that needs to be done. Pope Francis reminds us that technological power must be accompanied by responsibility, values, and conscience. The future of medical duty will be shaped in that shared space, where technology, foresight, and responsibility meet.³⁷