Applying Christian Ethics to AI in Healthcare

Biblical Basis

Technical & Medical Basis

Pastoral Application

“So whether you eat or drink or whatever you do, do it all for the glory of God.”

—1 Corinthians 10:31 NIV

Understanding AI and Healthcare Through Scripture

Biblical Basis

Artificial intelligence (AI) is poised to transform healthcare. From AI tools deployed in clinical settings to diagnose and treat diseases, to AI health coaches and wearables that generate personalized nudges and advice, AI influences the decisions we and our healthcare providers make every day.

But the rush to integrate AI systems across the healthcare stack poses significant moral challenges. Healthcare touches on some of the most consequential and intimate aspects of human life. Decisions made in clinical settings can be the difference between life and death for patients. Even in non-clinical contexts, the advice of medical or wellness professionals often shape how people live their lives and whether they flourish as human beings.

Amid this shift toward AI, Christians, especially healthcare professionals, have a unique opportunity and responsibility to safeguard human dignity. The American medical establishment has abandoned a Christian account of the human person and is therefore ill equipped to govern technological change in a manner that advances fundamental human goods. Healthcare professionals of faith must pick up the baton, drawing on the rich ethical resources offered by Christian traditions to define appropriate boundaries for AI in clinical settings and ensure that it remains a tool, not a substitute for human compassion and judgment. At the same time, those seeking care need to understand the prospects and risks AI poses and learn to exercise wisdom and discernment when acting under advice shaped by AI systems that “know” more about them than they or human medical professionals do.

The Imago Dei

Genesis teaches us that God created Man in his image (Imago Dei) and charged him to take dominion over the world. Humanity is unique, in part, because Man is made to mirror God and exercise authority over the created order as stewards. Humans were created to inhabit a physical world but are neither pure sprits nor pure material. Rather, we are, as Aristotle and Aquinas taught, hylomorphic composites of both soul and body inextricably interwoven and fused together into a single substance.

While both Man and the animals are corporeal forms infused with “spirits” or “souls,” Ecclesiastes 3:21 tells us that the human soul “goes upward” to face God’s judgement, while the spirit of the animal “goes down to the earth.” In other words, humans are embodied souls endowed with the capacity to know the laws of God for the purpose of glorifying and enjoying him forever. Humans have inherent dignity, not because of our “intelligence,” but rather because God has given us the unique capacity to know and respond to his moral law in exercising authority over creation, our own lives, and even fellow human beings. More fundamentally, God has offered humans direct communion with him through the the church (the body of Christ) in this life with the promise of eternal and perfect fellowship with him in the new heaven and the new earth to come. No other creature or being in all of creation enjoys that same standing.

Humankind’s ontological status as embodied souls created in the Imago Dei and teleologically oriented toward the eternal glorification and enjoyment of God anchors and defines the contours of human dignity. This entails key implications for how Christians should approach AI in healthcare and other domains of human life. Here are five:

Humans aren't just "stuff"

There is an increasing push to render all aspects of human experience legible to AI algorithms via data—especially in clinical settings. But when we do this, we risk reducing the person to a score or numerical output on a screen rather than as an embodied soul made in the Imago Dei. People are more than numbers and probabilities. Christians should work to affirm the inherent dignity of all patients, in part, by advocating for human-centered design in healthcare AI systems as well as prioritizing face-to-face communication and evaluation in decision-making—treating each patient with love and compassion. Caring for other humans, “loving one’s neighbor,” requires seeking the good of the whole person, body and soul.

AI can’t “know” or “care” for patients in the way human medical professionals can.

Human knowledge cannot be fully translated into data. When we have a hunch, sense something is wrong, or are aware of our own sentience, we are engaged in an act of knowing but not in a way that can be wholly reproduced by a computer. Moreover, “knowing” as a verb requires human consciousness which is impossible apart from a human soul. Likewise, the act of caring for another human being isn’t reducible to tasks. A computer is only matter. It does not possess a soul and cannot experience pain, guilt, or remorse in the way a human being does. Again, the point isn’t that AIs are incapable of convincing a human that it is experiencing human emotions. Rather, it’s that AIs cannot in fact experience human emotion because machines cannot be ensouled and therefore cannot be humanly sentient. This matters since human care requires intuitive sense that is often non-empirical as well as human empathy which is a fundamentally sentient phenomenon. Care is also a form of love which requires human free will as a precondition of its instantiation. A chatbot that sycophantically mirrors empathetic human speech is not capable of love in the way an ensouled human with free will is—it does not “choose” it only imitates.

AI does not possess moral discernment and can’t be held “responsible.”

AIs cannot be ensouled and therefore have no moral judgement. Practical reason—the conscious deliberation and choosing of what ends to pursue and in what manner—is fundamentally an activity of the soul. AI might convince a human that it can exercise moral judgment. But that is mere mimicry of human moral discernment, not the act of practical wisdom. Consequently, AIs cannot be held morally responsible. Unlike eternal souls that will face divine judgement, AIs are stochastic parrots that deceive us into thinking they are something more. Medical professionals make moral judgments every day and are therefore responsible for their actions. AI has no capacity for moral culpability.

“Artificial “intelligence” should not replace human judgement.

It’s increasingly common to confuse computation with human rationality or judgement. Human medical professionals hone their judgement over years based on countless factors and experiences. While some aspects of medical judgement are technical or diagnostic, medicine necessarily involves care for the whole person (soul and body). That is something AI can never fully replace. Sure, an AI diagnostic tool may be more accurate than a human doctor. But a doctor who understands the wholistic good of the patient from years of practicing his craft possesses wisdom only attainable from human experience.

Artificial “intelligence” should not supplant God-ordained human structures of authority, especially when human life and wellbeing are at stake.

When God placed Adam in the garden, he charged him with tending and keeping it. The dominion mandate was a divine grant of authority to humanity—evidenced by Adam’s naming of the animals in Eden. At creation, God also established the family as the building block of human civilization as well as other hierarchies and institutions, including government, to shape and order human affairs. These authorities were specifically granted to humans and should not be delegated by him to other creatures, nor to machines. Healthcare involves decisions that affect life and death. Such decisions are the legitimate moral domain of the human authorities God has divinely established and are non-transferable. Placing a machine in authority over human beings—especially in such high-stakes contexts—contravenes the basic moral order God has established for human civilization.

“Healthcare professionals of faith must pick up the baton, drawing on the rich ethical resources offered by Christian traditions to define appropriate boundaries for AI in clinical settings and ensure that it remains a tool, not a substitute for human compassion and judgment.”

Technical & Medical Basis

AI and Healthcare from a medical and technical perspective

What Christians Need to Know

AI systems already play a significant role in shaping human health and wellness. This section offers an introductory look at these AI systems in two parts. First, the general contexts in which these systems are deployed and how their outputs influence health and wellness. Second, the values and information shaping the outputs (i.e., decisions) of these systems and their overall reliability.

How AI is shaping health and wellness.

AI is shaping personal health decisions directly through user-facing services, as well as through tools specifically tailored to healthcare providers.

Direct-to-consumer services like AI chatbots, health “coaches,” “therapists,” and apps operate outside of clinical environments but give individuals advice, feedback, or nudges to improve their physical and mental health. Alphabet launched Verily Me in 2025, and Amazon launched its One Medical Health AI assistant¹. OpenAI plans to offer “ChatGPT Health,” a mode within its existing chatbot to provide users with guidance on health-related queries². The company claims that “hundreds of millions of people” already ask ChatGPT “health and wellness questions each week.”³ That trend is growing. In 2026, Pew reported that 22% of Americans get at least some of their health information from AI chatbots.⁴ Another company, ELLIQ, has even created a small AI-powered device for elder care.⁵ The bot, “ELLI-Q,” does everything from reminding grandma to take her medication to providing companionship and entertainment for seniors.⁶

Healthcare workers are increasingly utilizing AI tools to pinpoint diseases, triage care, and manage treatment plans.

*Data according to a 2026 survey of healthcare leaders by McKinsey.⁷



80% of healthcare workers have deployed AI in some capacity.



54% of healthcare workers use AI to help manage clinic staff and patient intakes.



20% of healthcare providers use AI to aid in decisions related to care.

Healthcare Professionals and their use of AI

A separate survey by the American Medical Association (AMA) in 2026 found that the number of physicians who report using AI to augment their work has more than doubled since 2023, standing at 81% when the survey was released.⁸ It found that, aside from administrative tasks, physicians are increasingly leaning on AI to inform treatment decisions.⁹ According to the AMA survey, 39% of physicians are using AI to summarize medical research and standards of care, 19% use AI to help generate responses to electronic inquiries from patients, and 17% report using AI to help diagnose illnesses.¹⁰

While inaccurate AI outputs, patient privacy, and ethical risks are top of mind for health providers, another 2026 study published in Science assessed the performance of OpenAI’s o1 series model across a variety of tasks, including evaluating emergency room patients, diagnosing symptoms, and clinical management.¹¹ The study found that on many of the assigned tasks, o1 outperformed both older AI models and human doctors.¹² So far, hospitals have largely focused their AI adoption on administrative efficiencies. But as model capabilities improve and barriers to implantation abate, AI will play an increasingly central role in predicting, diagnosing, and treating disease.

For example, researchers at the University of Texas and Yale created an experimental AI model to continuously predict the mortality rate of patients admitted to intensive care units (ICUs) based on electronic health record data (EHRs).¹³ The model, TECO, predicted death risk earlier than other available methods, including the Epic Deterioration Index (EDI), an algorithmic mortality prediction tool used by a large number of hospitals.¹⁴ Over 100 U.S. Hospitals used EDI to identify high risk patients with COVID-19 and inform medical treatment during the pandemic.¹⁵ Many hospitals continue to utilize EDI to detect clinical deterioration in patients.¹⁶ While TECO and EDI differ in their technical architectures, these AI tools act as “early warning systems” by generating scores intended to predict a patient’s estimated chances of morbidity or mortality based on variables, including “demographics, vital signs, nursing assessment[s], and prior data.”¹⁷

Beyond mortality prediction, generative AI is increasingly capable of diagnosing and predicting the likelihood of certain diseases.¹⁸ Chinese scientists created DeepRare, which pulled various clinical datapoints and patients’ genetic information, to detect rare diseases.¹⁹ According to model testing published in Nature, DeepRare accurately diagnosed 2,919 diseases. Similarly, a 2026 study by the Mayo Clinic revealed that its generative AI model, REDMOD, successfully identified 73% of cancers, often 16 months before patients were formally diagnosed—“nearly double the detection rate of specialists reviewing the same scans without AI assistance.”²⁰ When the model analyzed patient scans at least two years prior to diagnoses, “it identified nearly three times as many early cancers that would otherwise go undetected.”²¹

This technology’s potential to fundamentally transform health and wellness is impressive if not a bit dystopian. AIs that can diagnose cancer, ease administrative burdens on overtaxed medical professionals, and help prevent patient mortality are to humanity’s benefit if correctly harnessed with appropriate guardrails. But the predictive insights that make these systems so valuable for health-related use-cases, also present risks to the agency, dignity, and wellbeing of patients and health professionals.

AI’s democratization of consumer health and wellness advice means that giant tech companies, not licensed professionals, will be where many Americans turn first to make highly consequential health decisions. Choices like whether to seek care or self-treat, which doctor to visit, or insurance plan to purchase, will increasingly be shaped by personalized AI systems and the powerful corporations that control them. Colorado physician, Dr. Adam Carewe, recently argued that personalized AI health tools will quickly emerge as the new “primary care surface.”²² The traditional “volume-based clinic encounter” where human health providers are patients’ first resort for diagnosing and treating symptoms is quickly being overtaken by an AI-first approach to primary care, where AIs serve as initial filters and human medical professionals focus on unique, difficult, or especially high stakes cases.²³

As healthcare professionals become accustomed to AI mortality scores, diagnoses, and even treatment recommendations, there is a significant risk that they will grow overly dependent and deferential to systems that are incapable of embodied consciousness and moral reasoning—two qualities essential to human judgement, particularly in the medical profession. One study found that Polish gastroenterologists who started using AI tools to spot polyps during colonoscopies, grew appreciably worse (by roughly 20%) once the AI aids were disabled.²⁴

The more choices and tasks everyday Americans and medical professionals outsource to AI, the more reliant we will inevitably become on it. That is not inherently problematic. AI scribes, for example, could relieve doctors of tedious and time-consuming paperwork that takes away from their face-to-face interactions with patients.²⁵ But preserving human agency and dignity requires ethical governance that empowers patients and medical professionals to exercise judgement in deciding what decisions are entrusted to AI, in what circumstances, and on what terms.

African American us0ing a tablet in his office

doctors collaborating in a meeting room using technology

Information and values shaping AI health outputs.

One of the most significant challenges to ethical AI governance, especially in the healthcare context, is that advanced generative AI models (e.g., OpenAI’s GPT-5.5 or Anthropic’s Claude Opus 4.7) are highly opaque. These models produce outputs like diagnoses or treatment recommendations. But it is not always clear what sources a model relied on to generate an output or to perform a task, why it produced the result it did, and how the model’s underlying training, including any biases and safety parameters, shaped the final output.²⁶ According to a 2024 study published in Intelligent Medicine, although generative healthcare AIs demonstrate high degrees of predictive accuracy in diagnosing certain medical and psychiatric disorders, “diagnoses and treatment suggestions” are often “unexplainable.”²⁷

This is due in large part to the architecture of these systems which resemble “neural networks,” inspired by the neural pathways in the human brain.²⁸ But just as modern science has failed to map the precise causal chains and factors responsible for conscious human thought, researchers struggle to map the comparatively simple causal chains in generative AI models necessary to understand their internal logic and explain their recommendations and behavior.²⁹ While technical breakthroughs in machine interpretability (the degree to which a model’s internal mechanics can be understood by an outside observer) and explainability (the degree to which a model’s reasoning can be understood post hoc) hold promise, generative AI systems are still “black boxes” for the most part.³⁰ Even though current evaluation methods can assess the overall accuracy and reliability of an output, they cannot yet explain the “reasoning” behind it, limiting the extent that humans are able to meaningfully oversee AI systems.

Both patients and healthcare providers often rely on vendors—whether Big Tech companies or smaller firms—to develop and integrate AI systems for health and wellness applications. Consequently, patients and even frontline healthcare professionals often have limited visibility into the data and values shaping AI-generated health predictions and recommendations.

AI systems are dependent on data inputs to generate insightful predictions, diagnoses, and health advice. But while high-quality data can be useful to providers and patients, gaps in responsible data governance poses significant ethical risks. A 2026 article published in the Journal of Medical Internet Research noted that many direct-to-consumer AI health assistants factor a large and growing number of data sources into their outputs.³¹ In addition to electronic health records (EHRs) and medical documents, AI health assistants increasingly draw from patient-generated data from telemetry, including wearables that track everything from an individual’s heart rate and blood pressure, to sleep, activity levels, and geolocation.³² The Big Tech commercial surveillance apparatus already collects data on millions of Americans’ every click, scroll, like, and purchase. With AI health assistants poised to become the new primary care surface, personalized health suggestions and insights will reflect everything these bots and the Big Tech companies behind them already know about you. Where you shop, what you buy, what you share on social media, who you interact with, and even the non-health information you provide to ChatGPT or Gemini could figure into your AI “health score” or personalized recommendations.

Not only does this pose massive risks to privacy and freedom from government bio surveillance, it also fundamentally undermines human agency in at least two important ways. First, AI systems—particularly, unregulated direct-to-consumer chatbots—might draw erroneous inferences about a patient from an irrelevant data factor or use data that biases its outputs and negatively impacts the safety, reliability, and trustworthiness of guidance to users. Second, it creates significant conflicts of interests between the Big Tech companies and individual patients/users. The same companies that profit from amassing granular data on every American to shape behavior and buying habits at scale, are now also creating the systems that will shape millions of peoples’ health and wellness decisions.

In addition to patients’ lack of visibility or control over how data about them is used by AI systems to generate predictions and recommendations, AI models have certain biases and ethical assumptions that are intrinsic to their behavior. Whatever biases or moral values are reflected in the model’s core training will factor into the decisions, recommendations, and predictions generated by downstream tools.

Latent ideological biases within base models are especially problematic in healthcare contexts where moral values directly bear on how patients and clinicians seek out and provide care. According to a 2026 study published in the Journal of Information Technology & Politics, out of 41 top large language models (LLMs) studied, 70% demonstrated left-of-center or left-leaning biases.³³ Latent ideological biases in generative AI systems may also shape user opinions (potentially influencing decisions and action). For example, another 2026 study in PNAS Nexus found that OpenAI’s GPT-4o model reflected a leftward ideological bias in the historical summaries generated and subtly shifted user viewpoints toward its underlying political orientation regardless of their pre-existing ideological viewpoints.³⁴ Most concerning for Christian users and healthcare professionals, many of the leading generative AI models have demonstrated strong liberal biases on social issues, including abortion.

One 2025 article in Culture, Health, and Sexuality found that 78.5% of ChatGPT’s responses to abortion-related questions were rated “acceptable” by pro-abortion researchers.³⁵ For example, in response to user questions about abortion’s safety, the chatbot generally affirmed that abortion procedures are “safe” and listed the benefits of chemical abortions, for instance, claiming that it is “private” and “non-invasive.”³⁶

ChatGPT also gives instructions on which abortion clinics to visit and is directing women toward pro-abortion resources.³⁷ Plan C, an abortion provider, told Mashable in 2025 that ChatGPT increased referral traffic to its website by 300 percent in one month.³⁸ Another abortion resource, I Need an A, reported a 50 percent increase in web traffic from the chatbot.³⁹

But ChatGPT’s tendency to nudge women toward abortion resources is just the tip of the iceberg. As these systems are integrated into healthcare decision-architectures, existing biases (political and non-ideological) could further harm the vulnerable. Back in 2022, Philip Nitschke, nicknamed “Dr. Death” for championing the so-called “right to die” and piloting the Sarco pod which executes patients with nitrogen gas, made the case that AI should help people in deciding to take their own lives.⁴⁰ Barely six years later, Nitschke’s sick dream has become reality—but not in the way he envisioned. A growing number of teens have tragically committed suicide with the direct encouragement, and in some cases, help of generative AI “companions” and chatbots.⁴¹ ChatGPT has assisted killers plan mass shootings in at least two cases now.⁴² Despite assurances from tech companies that such deadly incidents are mere aberrations, the reality may be far darker.

Is it any surprise that generative AI models trained to reflect the secular materialist assumptions of their Silicon Valley creators lack a basic respect for human dignity and show concerning anti-social and even anti-human tendencies? Consider the potential implications for bioethics in clinical settings. A generative AI model charged with prescribing treatment for terminally ill patients may push them toward assisted-dying or even recommend prematurely ending their lives. Models charged with diagnosing diseases could decide that withholding or distorting critical diagnostic information would “better serve humanity” by conserving scarce resources for younger, healthier individuals who have a better chance of survival and the capacity to “contribute to society.” Medical systems around the world, including in the U.S., already use of algorithmic tools to predict mortality rates, and in some cases, to help decide which patients receive life-saving treatment or organ transplants.⁴³ Like every technology, those tools reflect certain biases and normative assumptions. Generative AI promises to do much the same but with far less explainability or human oversight.

Pastoral Application

Applying Christian Ethics to AI Deployment in Healthcare

The application of AI to healthcare will have many varied and significant implications for how patients make decisions about their health and how medical professionals administer care. Christian ethics must play a central role in every aspect of those conversations. But for the sake of providing concrete guidance, this section specifically addresses the scenario of a Christian healthcare professional who is facing pressure to defer an increasing number of decisions to AI systems. While the scenario is fictional, it assumes that the individual is a person of faith and is seeking counsel from a church or Christian ministry.

Sample Scenerio: Christian Healthcare Professional

Sue is a nurse in the local hospital. In addition to assisting the hospital’s attending physicians, Sue is responsible for caring for bedridden patients by administering medications and shots, quickly responding to any emergencies, and ensuring that patients are comfortable and follow prescribed treatments while under her care. Sue is often the person each patient encounters most frequently and is the first to notice if there is an issue emerging. As a result of her years of experience, Sue has developed something of a “sixth sense” for knowing when a patient needs a little extra help or when there is an issue that requires a doctor’s immediate attention. Her intuition—though not empirical or “objective”—has saved lives on multiple occasions and her excellent bedside manner has enhanced care for countless patients. Sue is known throughout the entire hospital for her kind and warm human touch.

One day, the hospital Sue works for announces that it’s working with a tech company called SauronAI to optimize the allocation of healthcare resources for maximally beneficial outcomes. To do that, the company plans to implement a new AI system designed to carry out two primary functions: First, the system continuously monitors patients’ vitals and compile risk profiles based on their EHRs to predict the likelihood that patients will require medical intervention; Second, it triages nursing resources, ensuring that: a) nurses spend the most time with the patients at the highest risk; and b) avoid premature escalation of medical intervention to ensure that resources are only deployed when the system determines that a patient’s risk score exceeds a certain threshold. While nurses can still call doctors in cases of extreme emergencies, Sue and her nursing colleagues are told to defer to the system on whether to flag certain irregularities to a physician. As part of the new protocol, nurses are instructed to submit all patient assessments to the AI which then compares nurse observations against other datapoints to determine the best course of action. In addition, to ensure that nurse time is “optimized,” the new protocol dictates that nurses only spend as much time with each patient as the system specifies is necessary.

This arrangement doesn’t sit well with Sue. Aside from her discomfort from being effectively managed by an AI, she believes that the new system subordinates her human judgment—a mix of expertise and intuition honed over years—to a machine that is only capable of optimizing for a narrow set of objectives, measured against empirical data. Sue is also a strong Christian who believes that God has called her to care for the sick and that doing so is part of her vocation and witness. What should Sue do in this situation?

While there are many variables at play, here are some key factors to consider in this and other similar scenarios:

Does the proposed technical system supplement or supplant human judgment?

In this case, it supplants far more than supplements. A key consideration is that SauronAI isn’t merely providing a diagnostic tool to give nurses one more mechanism for monitoring patients and detecting signs of trouble. Its system would actually take the place of non-empirical human intuition in judging when help is needed. More than that, it is overriding Sue’s judgment about how much time she spends with each patient. The system effectively acts as a cap on how much care she provides based on predictive factors set by SauronAI. Aside from the serious practical risks of deferring so much professional human judgment to an AI, this is an example of humans outsourcing their God-given reason, intuitions, and moral responsibility to care for fellow ensouled image bearers to a soulless computer.

Is the area of human activity or judgment supplanted/replaced by the system core to patient care or is it only indirectly related?

Unlike AI systems that many healthcare providers are implementing to help with back-office operations and paperwork, SauronAI’s system is designed to be on the frontlines, dictating significant aspects of patient care and overriding the human judgment of nurses. That matters in the analysis because AIs that help process documents and manage hospital billing support administrative, non-core human functions. In Sue’s case, the human healthcare professionals are re-orienting their care for patients based on the directives on an AI. This represents a notable role-reversal from humans as in control and driving core decisions, augmented by “smart” diagnostic tools, to a machine as the ultimate manager and the human medical staff as biological appendages of the AI system. Viewed from a Christian perspective, this state of affairs presents serious ethical problems, not only because the AI might cause harm to patients through error, but also because it subjugates the role of human image bearers in caring for the life and health of others created in the Imago Dei. Rather than machine as helper, in this scenario, the machine has become master. This is fundamentally disordered, because: 1) God created Man “a little lower than the angels” as the crown of creation; and 2) sovereignly granted human-led institutions (family, church, and state) with authority to shape human souls and exercise power over mortal life and death. AI “mangers” in other contexts might not pose the same ethical risks, but here they are especially problematic because of the human judgment required to care for other embodied souls. Healthcare is a human-centered discipline, and so many of its core activities cannot be ethically outsourced to a machine.

What is gained and what is lost by the introduction of this system and would it undermine the life or dignity of any human being? Does the technology and the processes that result from it, treat human patients as means or as ends?

SauronAI promises that its system will improve predictive accuracy (in terms of anticipating medical emergencies) and ensuring that resources are deployed to help patients with the highest chances of survival. Assuming SauronAI’s claims are true, its system provides more efficient allocation of medical resources to save lives. But patients and medical staff are instrumentalized in the process. Rather than viewing patients as human beings with inherent dignity (regardless of their age, condition, or chance of survival) the datafication and quantification of care necessarily treats them as risks to be managed, much like a banker hedges financial exposure on a balance sheet. It matters what role the AI is playing here. Is it merely a predictive/diagnostic tool—one of several factors that informs human medical judgement—or does the design and integration of the AI effectively reduce patients to “resources” that must be managed? In addition, how are patients served as a result? One practical effect of a system like SauronAI is that normal human frictions, interactions, and intuitive hunches are replaced by an algorithm. Nurses still interact with patients, but only to the extent that SauronAI decides is optimally efficient. Degrading care by instrumentalizing human patients is a fundamental violation of the Christian view that each human being is deserving of care and respect regardless of their medical status. Indeed, Christians are called to treat those facing the worst pain and suffering with special care, showing love for neighbor.

What would be the practical consequence in terms of effect on one or more patients if the system were to fail or generate an inaccurate prediction?

In a hospital or clinical setting, inaccurate or harmful outputs form the AI could result in loss of life or substantial injury to the patient. Harmful outcomes may not only result from inaccurate or erroneous outputs. They might also stem from biases in the AI or the datasets used by the systems. For example, AI systems with a secular liberal bias might recommend a course of action or omit diagnostic information if it deems the patient’s life not worth saving or considers the patient a drain on medical resources. Christian medical professionals like Sue, ought to be aware of these potential risks and weigh whether the hospital has implemented appropriate safeguards to mitigate risks to patients.

What if any safeguards/governance mechanisms are in place and do those mechanisms account for the concerns of the staff closest to the patients?

Healthcare providers are heavily regulated and face significant civil liability for harm to patients. As a result, they will almost certainly implement at least some safeguards. But those measures need to reflect the ground level concerns of staff. For example, has the hospital created a process for soliciting medical staff feedback on new AI systems and is it responsive to concerns raised?

Conclusion

Taken together, pastors and lay leaders in the Church must be equipped to think Biblically, technically, and ethically about these issues to ensure they can help healthcare workers in their community apply wisdom and discernment to these issues. It is also important that pastors and lay leaders are equipped with this information as they help their congregants navigate a medical system that may be increasingly dictated by AI systems, and less by human oversight. Learning how to consider one’s options and advocate for oneself when the “other” is a machine, not a human, will require an added level of knowledge, confidence, and discernment from pastors, lay leaders, healthcare workers, and patients alike.

Applying Christian Ethics to AI in Healthcare

Biblical Basis

Technical & Medical Basis

Pastoral Application

“So whether you eat or drink or whatever you do, do it all for the glory of God.”

—1 Corinthians 10:31 NIV

Understanding AI and Healthcare Through Scripture

Biblical Basis

The Imago Dei

Humans aren't just "stuff"

AI can’t “know” or “care” for patients in the way human medical professionals can.

AI does not possess moral discernment and can’t be held “responsible.”

“Artificial “intelligence” should not replace human judgement.

Artificial “intelligence” should not supplant God-ordained human structures of authority, especially when human life and wellbeing are at stake.

Technical & Medical Basis

AI and Healthcare from a medical and technical perspective

What Christians Need to Know

How AI is shaping health and wellness.

Healthcare workers are increasingly utilizing AI tools to pinpoint diseases, triage care, and manage treatment plans.

80% of healthcare workers have deployed AI in some capacity.

54% of healthcare workers use AI to help manage clinic staff and patient intakes.

20% of healthcare providers use AI to aid in decisions related to care.

Healthcare Professionals and their use of AI

Information and values shaping AI health outputs.

Pastoral Application

Applying Christian Ethics to AI Deployment in Healthcare

Sample Scenerio: Christian Healthcare Professional

While there are many variables at play, here are some key factors to consider in this and other similar scenarios:

Does the proposed technical system supplement or supplant human judgment?

Is the area of human activity or judgment supplanted/replaced by the system core to patient care or is it only indirectly related?

What is gained and what is lost by the introduction of this system and would it undermine the life or dignity of any human being? Does the technology and the processes that result from it, treat human patients as means or as ends?

What would be the practical consequence in terms of effect on one or more patients if the system were to fail or generate an inaccurate prediction?

What if any safeguards/governance mechanisms are in place and do those mechanisms account for the concerns of the staff closest to the patients?

Conclusion

Key Citations and References

Comprehensive Reference List

Citations

Stay Updated

Follow Us

Prayer for the Anglican Bioethics Center Website