INTRODUCTION
In the annals of human history, few
innovations have been as transformative
as the printing press and the machinery
of mass production in the Industrial
Revolution. Today, we stand on the
precipice of another monumental shift:
the rise of generative artificial intelligence
(gAI). This technological marvel, with
its potential to revolutionize sectors like
Catholic health care, beckons us to “read
the signs of the times and respond to them
in light of the Gospel.”1
The complexities
of gAI are vast, and our aim is to offer some
practical ethical guidance on this immense
topic. Catholic moral and social teaching
has insights that can guide the development,
deployment, and evaluation of artificial
intelligence in health care. This essay
argues that Catholic social teaching (CST)
in particular offers insights that shape
axioms for generative AI in health care
applications.
WHAT IS GENERATIVE AI?
AI, or Artificial Intelligence, refers to the
simulation of human intelligence in machines.
It's a branch of computer science aiming to
create systems able to perform tasks that usually
require human cognition, such as decision-making, pattern recognition, understanding
language, and problem-solving. When AI
is qualified as “generative,” it means that
it is a type of artificial intelligence that has
the capacity to produce outputs that mimic
human-created content (text, images, etc.).
At its core, AI is a tool. Just as a hammer is
an extension of the hand for driving nails,
AI can be seen as an extension of human
cognition for processing data, recognizing
patterns, and performing tasks. One may
think of AI as a highly advanced cognitive-arithmetic-linguistic-algorithmic tool: AI has
multifaceted capabilities from basic arithmetic
to complex cognitive tasks like natural language
understanding.
Another way to conceptualize artificial
intelligence is through its three primary
functions: automation, augmentation, and
assistance. First, AI's automation capabilities
transfer specific tasks from human oversight to
algorithmic control, thereby reducing manual
labor and increasing efficiency. Second, AI
serves to augment human capabilities by
enhancing cognitive processes and expanding
the collective knowledge base, thereby enabling
more informed decision-making and problem-solving. Lastly, AI assists in streamlining
operations by providing real-time support and
guidance, which in turn lightens the human
workload and improves overall productivity.
However, the line can get blurred because
of AI's ability to mimic certain human-like
qualities, such as conversation, generating
digital images, or playing games. Regardless
of how advanced or "intelligent" an AI might
seem, it does not have feelings, consciousness,
or self-awareness. It operates based on the code
and algorithms it's been designed with, making
it a tool created and directed by humans.
There is a parallel between the rise of AI and
the era of industrialization. The following
five points suggest this: One, with AI, many
anticipate the displacement of jobs. Just as
industrialization led to the automation of
many manual, labor-intensive tasks (making
certain crafts obsolete), AI has the potential to
automate many “thinking-intensive” jobs that
involve data analysis, customer service, and
even some aspects of decision-making.
Two, relatedly, AI could require a skill shift.
Industrialization required workers to acquire
new skills to operate machines. Similarly, the AI
era requires a workforce that understands how to work with, manage, and even program these
new tools. At the time of this writing, a great
many early adopters of AI have focused on the
art of the prompt: how to get AI like ChatGPT
to produce the intended results the user wants
by making the prompt “perfect.”
Three, many expect AI to increase productivity
of certain work. Just as machines increased the
scale and efficiency of production of goods, AI
can increase efficiency in various sectors, from
finance to healthcare, by handling large datasets
and performing complex calculations at speeds
unimaginable to humans.
Four, many see immense societal implications
of AI. The industrial era brought about
significant societal changes, from urbanization
to changes in work-life balance. Similarly, AI
has the potential to bring profound societal
shifts, such as changes in how we view privacy,
the nature of work, or even what tasks are
deemed valuable.
Finally, there are ethical concerns with AI.
Industrialization raised concerns about worker
safety, fair wages, and working conditions. AI
introduces its own set of ethical issues, from
bias in algorithms to surveillance concerns.
Just as industrialization transformed societies,
economies, and ways of life, AI promises (or
threatens, depending on one's perspective)
to bring about its own set of transformative
changes. The challenge for societies is to
harness the benefits while mitigating potential
drawbacks and ensuring equitable outcomes.
We may think of AI, and generative AI in
particular, as marking the industrialization of
thought.
This term implies the mechanization and
systematic production and execution of
cognitive and creative tasks, similar to how
industrialization referred to the mechanization
of physical labor. For example:
Standardization and Scalability: Just as
industrialization led to the standardized
production of goods on a large scale, AI
allows for the standardized processing of data
and decision-making on scales previously
unattainable. An AI model, once trained, can
be deployed countless times across different
devices and platforms, producing consistent
results.
Efficiency and Speed: Industrial machines
increased the speed of production. Similarly, AI
can process and analyze vast amounts of data at
speeds far surpassing human capabilities.
Specialization: With industrialization,
machines were often designed for specific tasks,
leading to specialized production lines. In AI,
there are specialized models for various tasks,
from image recognition to natural language
processing.
Transformation of Human Labor: Just as
machines reduced the need for manual labor,
AI reduces the need for human cognitive labor
in certain areas. Tasks like data analysis, which
might take humans hours, can be completed in
moments by AI.
Depersonalization: A criticism of
industrialization was that it could lead to the
depersonalization of work, turning craftsmen
into mere cogs in a machine. Similarly, there's
a concern that relying too heavily on AI,
especially in areas like decision-making, might strip away the human touch, intuition, or
ethical considerations.
Transformation of Skill Sets: As
industrialization changed the skills workers
needed, AI's rise emphasizes the need for new
skills in the modern workforce, such as data
literacy and understanding AI ethics.
In many ways, the phrase industrialization
of thought aptly captures the transformative
impact of AI on cognitive tasks and broader
societal functions. However, it is essential
to recognize that while AI can simulate
many aspects of human thought, it lacks
consciousness, emotions, and the nuanced
understanding that humans bring to tasks.
Generative artificial intelligence (gAI)
represents a significant shift in the realm of
computational capabilities. Unlike traditional
AI systems that primarily focus on analysis
and prediction, gAI is designed to create.
This creation can range from generating
coherent text to simulating intricate biological
processes. The potential applications of gAI
are vast, especially in sectors like health
care. For instance, gAI can revolutionize
diagnostics by analyzing extensive datasets to
identify patterns that might be imperceptible
to the human eye. This could lead to the
early detection of ailments even before they
manifest. Additionally, by understanding a
patient's unique genetic makeup, lifestyle, and
medical history, gAI can offer personalized
care, optimizing treatment outcomes. Another
promising application is in the realm of
research acceleration. The drug discovery
process, which traditionally spans several years,
could be significantly condensed with gAI
simulating molecular interactions, predicting drug efficacy, and ensuring safety.2
Sara Vaezy explains four strategic domains
of gAI applications in health care.3
First, in
the clinical domain, gAI has the potential to
support clinical decision-making, automate
mundane tasks, and assist providers with
documentation. Second, from the patient’s
perspective, gAI could augment patient
experience in a highly personalized, precise
way based on their unique needs, motivations,
preferences, and history. Third, the
administrative domain contains numerous
gAI opportunities for various tasks such as
predictive scheduling, billing applications, etc.
Finally, Vaezy points to several gAI applications
to back-office functions such as applications
that intercept and redirect inquiries to the best
channel or outlet to support the specific needs
of the patient or consumer.
Generative AI technologies are rapidly maturing
and finding applications in various domains,
including software engineering. For instance,
gAI can be used in software engineering use
cases such as translating natural language
to code, code translation, and code autocompletion.4
However, the introduction of gAI
into various sectors also brings forth a plethora
of ethical considerations. Concerns range from
potential infringements on copyrights due to
the replication and production of content by
gAI, the risk of job losses due to automation,
to challenges in discerning truth from fiction
given the ability of AI to create realistic
content.5
The ethical implications surrounding generative
AI are profound. While the technology offers
promising advancements, it also underscores
the multifaceted ethical landscape that demands careful consideration and proactive measures as
it continues to evolve.6
As the above suggests, the issues are vast and
many:
- Theft of Intellectual Property: Generative AI
can replicate and produce content, leading
to potential infringements on copyrights
and the devaluation of original creations.
- Displacement of Workers: As AI automates
tasks, there's a risk of job losses, especially
in sectors reliant on repetitive tasks,
potentially leading to economic and social
disruptions.
- Loss of Autonomy: Over-reliance on AI
recommendations can diminish human
decision-making, making individuals overly
dependent on algorithms for choices.
- Erosion of Human Dignity and Dignity
of Work: Beyond automating tasks, AI
can reduce the perceived intrinsic worth
of human contributions, undermining the
unique value and experiences individuals
bring.
- Data Privacy and Confidentiality: AI
models, especially those that generate
content based on vast datasets, can
inadvertently reveal private information or
patterns, posing risks to individual privacy.
- Bias and Discrimination: AI models can
reflect and amplify societal biases present
in their training data, leading to unfair or
discriminatory outputs.
- Authenticity and Truth: The ability of AI
to create realistic content, like deepfakes,
challenges our ability to discern truth from
fiction, potentially enabling misinformation.
- Economic Inequality: The concentration
of AI capabilities among a few entities
can exacerbate economic disparities, with wealth and power becoming more
centralized.
- Safety and Reliability: Advanced AI models
can produce unpredictable results, posing
risks when deployed in critical sectors.
- Depersonalization: An over-dependence
on AI for personal tasks can diminish
human-to-human interactions, leading to
impersonal and detached experiences.
- Transparency and Accountability:
The "black box" nature of some AI
models can obscure decision-making
processes, challenging accountability and
understanding.
- Environmental Concerns: The
computational demands of training AI
models can lead to significant energy
consumption, raising environmental and
sustainability concerns.
- Over-reliance and Loss of Skills: Excessive
dependence on AI can result in the atrophy
of essential human skills, as machines take
over tasks previously done by humans.
- Anthropological Implications: AI's ability
to create art or philosophical content
raises questions about human uniqueness,
creativity, and consciousness.
- Regulatory and Legal Challenges: The rapid
advancement of AI can outpace legal and
regulatory frameworks, leading to potential
conflicts and ambiguities.
Each of these points underscores the
multifaceted ethical landscape of generative AI,
emphasizing the need for careful consideration
and proactive measures as the technology
evolves. The ethical implications are profound,
and the Catholic Social Teaching (CST) offers
a beacon.
PILLARS OF CATHOLIC SOCIAL TEACHING
CST, rooted in millennia of theological
reflection, provides a moral compass. Drawing
from many documents of modern CST, we
find that a number of key values, ends, and
mechanisms to effectuate change. Each of these,
in turn, point to general ethical principles
guiding Catholic health care. I will organize
our reflections of CST around three pillars –
axiological, eschatological, and sociological – to
shed light on the key principles these bring to
bear on generative AI questions.
Axiological Pillar: The axiological pillar describes
essential values at the center of human
personhood, communal living, and relationship
with God. These are human dignity, the
common good, and stewardship.
To begin, human dignity is the inherent
dignity rooted in being created in the image
and likeness of the divine: “Human persons
are willed by God; they are imprinted with
God's image. Their dignity does not come from
the work they do, but from the persons they
are.”7
As an essential value, human dignity
corresponds to two general principles: respect
human dignity and respect human life from
conception to death. These are interrelated, of
course, but distinct principles guiding behavior.
Next, the common good refers to the
context and capacity for human flourishing
in community. These words describe the
common good at Vatican II:
…the sum of those conditions of social
life which allow social groups and their
individual members relatively thorough
and ready access to their own fulfillment, today takes on an increasingly universal
complexion and consequently involves
rights and duties with respect to the whole
human race. Every social group must
take account of the needs and legitimate
aspirations of other groups, and even of
the general welfare of the entire human
family.8
Corresponding to the value of the common
good we have the general principles of
promoting the common good and enabling
participation in the common good itself.
A third essential value in CST I will name as
stewardship, which pertains to the dignity of
work: humankind’s participation in God’s plan
as created co-creators. St. John Paul II had this
to say:
Even though it bears the mark of a bonum
arduum, in the terminology of Saint
Thomas, this does not take away the fact
that, as such, it is a good thing for man.
It is not only good in the sense that it is
useful or something to enjoy; it is also
good as being something worthy, that is to
say, something that corresponds to man's
dignity, that expresses this dignity and
increases it. If one wishes to define more
clearly the ethical meaning of work, it is
this truth that one must particularly keep
in mind. Work is a good thing for man—a
good thing for his humanity—because
through work man not only transforms
nature, adapting it to his own needs, but
he also achieves fulfilment as a human
being and indeed, in a sense, becomes
'more a human being'.9
More recently, Pope Francis had this to say in his encyclical on caring for the Earth:
We were created with a vocation to work.
The goal should not be that technological
progress increasingly replace human work,
for this would be detrimental to humanity.
Work is a necessity, part of the meaning of
life on this earth, a path to growth, human
development and personal fulfilment.10
In sum, stewardship calls upon us to abide
by several general principles: (1) honor the
spirituality of work, (2) respect the dignity
of work itself and the workers (cf. above), (3)
prioritize the worker over utility and efficiency,
(4) exercise just use and allocation of resources
corresponding to the universal destiny of
goods (versus private property), and (5) act to
maximize sustainability of resources.
Eschatological Pillar: The eschatological pillar
orients humankind to the ends of God’s
invitation: a transcendent horizon fulfilled by
our love for God and for neighbor. This is a
‘now and not yet’ pillar. The horizon includes
ends such as responsibility and religious liberty,
social justice, integral human development, and
integral ecology.
The eschatological horizon in our tradition calls
for responsibility and religious liberty. In the
words of Dignitas Humanae:
In all his activity a man is bound to follow
his conscience in order that he may come
to God, the end and purpose of life. It
follows that he is not to be forced to act in
a manner contrary to his conscience. Nor,
on the other hand, is he to be restrained
from acting in accordance with his
conscience, especially in matters religious.
The reason is that the exercise of religion,
of its very nature, consists before all else
in those internal, voluntary and free acts
whereby man sets the course of his life
directly toward God. No merely human
power can either command or prohibit
acts of this kind. The social nature of man,
however, itself requires that he should
give external expression to his internal
acts of religion: that he should share with
others in matters religious; that he should
profess his religion in community. Injury
therefore is done to the human person and
to the very order established by God for
human life, if the free exercise of religion
is denied in society, provided just public
order is observed.11
Thus, in terms of general principles related to
responsibility and religious liberty, we have
the following. One, persons and corporations
should act responsibly and be held accountable.
Two, respect for personal and corporate
conscience should be established in law within
the parameters of the public order. Finally,
respect for diversity of views should be a
hallmark of communal living.
Social justice is another key component of our
eschatological horizon. Lisa Cahill defines
social justice as “inclusive participation in the
common good.”12 The Compendium of the
Social Doctrine of the Church states, “Ever
greater importance has been given to social
justice, which represents a real development
in general justice, the justice that regulates
social relationships according to the criterion of
observance of the law.”13
The general principles as they relate to social
justice include (a) promoting participation in society, (b) establishing commutative fairness
between parties, (c) encouraging contributive
fairness of individuals and groups, (d) ensuring
proper distribution of benefits and burdens,
and (e) exhibiting charity in the absence of
justice.
Next, the eschatological component includes
integral human development in our horizon.
Benedict XVI states,
The truth of development consists in its
completeness: if it does not involve the
whole man and every man, it is not true
development. This is the central message of
Populorum Progressio, valid for today and
for all time. Integral human development
on the natural plane, as a response
to a vocation from God the Creator,
demands self-fulfilment in a 'transcendent
humanism which gives [to man] his
greatest possible perfection: this is the
highest goal of personal development.' The
Christian vocation to this development
therefore applies to both the natural plane
and the supernatural plane; which is
why, 'when God is eclipsed, our ability to
recognize the natural order, purpose and
the ‘good' begins to wane.'"14
The general principles of integral human
development include but are not limited to
the following. One, design, development, and
deployment of technology should be in service
to the person, not vice versa. Two, persons
should be afforded the opportunity to develop
competencies and talents. Three, institutions
should cultivate an appreciation of the human
person in totality.
Finally, an integral ecology is a component of the eschatological pillar of CST. In Laudato Si’,
Pope Francis writes,
Since everything is closely interrelated, and
today’s problems call for a vision capable
of taking into account every aspect of
the global crisis, I suggest that we now
consider some elements of an integral
ecology, one which clearly respects its
human and social dimensions. [...] When
we speak of the 'environment', what we
really mean is a relationship existing
between nature and the society which
lives in it. Nature cannot be regarded as
something separate from ourselves or as
a mere setting in which we live. We are
part of nature, included in it and thus in
constant interaction with it.15
Ensuring that technology is not “severed” from
ethics, at least two general principle(s) apply:
(1) understand the interconnectedness of all
things, and (2) exercise sustainable development
and use of technology.
Sociological Pillar: The sociological pillar
provides specific mechanisms to be exercised in
social contexts in pursuit of the ends and values
mentioned above. First, solidarity, based on
the connection and relationship of humankind,
is “a firm and persevering determination to
commit oneself to the common good. That is
to say to the good of all and of each individual,
because we are all really responsible for all.”16
Thus, the general principles of solidarity
include (1) embracing a culture of encounter,
(2) exercising empathy, (3) build unity with
diversity, and (4) engage inclusive practices.
Second, subsidiarity ensures that decision-making should be localized, ensuring community relevance. Pius XI writes,
Just as it is gravely wrong to take from
individuals what they can accomplish
by their own initiative and industry and
give it to the community, so also it is an
injustice and at the same time a grave evil
and disturbance of right order to assign
to a greater and higher association what
lesser and subordinate organizations can
do. For every social activity ought of its
very nature to furnish help to the members
of the body social, and never destroy and
absorb them.17
General principles of subsidiarity include (1)
shift power to those more proximate to the
issues, (2) democratize technology and access
to it, (3) disclose information appropriately to
exercise due transparency with stakeholders,
and (4) obtain consent from appropriate parties
as possible.
Third, the preferential option for the poor or
marginalized entails concrete actions that are
always just and partial to those in need. From
the Church’s Compendium we learn,
The principle of the universal destination
of goods requires that the poor, the
marginalized and in all cases those
whose living conditions interfere with
their proper growth should be the focus
of particular concern. To this end, the
preferential option for the poor should
be reaffirmed in all its force. “This is
an option, or a special form of primacy
in the exercise of Christian charity, to
which the whole tradition of the Church
bears witness. It affects the life of each
Christian inasmuch as he or she seeks to imitate the life of Christ, but it applies
equally to our social responsibilities and
hence to our manner of living, and to the
logical decisions to be made concerning
the ownership and use of goods. Today,
furthermore, given the worldwide
dimension which the social question has
assumed, this love of preference for the
poor, and the decisions which it inspires
in us, cannot but embrace the immense
multitudes of the hungry, the needy, the
homeless, those without health care and,
above all, those without hope of a better
future.”18
For the preferential option of the poor, we see
the following general principles: (1) promote
health equity and equal opportunity, (2) invite
those marginalized to participate in design and
decision-making procedures.
Fourth, the sociological pillar includes corporal
works of mercy as a call to help those in need.
Again, the Compendium states,
The Church's love for the poor is inspired
by the Gospel of the Beatitudes, by the
poverty of Jesus and by his attention to
the poor. This love concerns material
poverty and also the numerous forms
of cultural and religious poverty. The
Church, “since her origin and in spite of
the failing of many of her members, has
not ceased to work for their relief, defence
and liberation through numerous works
of charity which remain indispensable
always and everywhere.” … [T]he Church
teaches that one should assist one's fellow
man in his various needs and fills the
human community with countless works
of corporal and spiritual mercy. … [E]ven if the practice of charity is not limited
to alms-giving but implies addressing
the social and political dimensions of
the problem of poverty. In her teaching
the Church constantly returns to this
relationship between charity and justice:
“When we attend to the needs of those in
want, we give them what is theirs, not ours.
More than performing works of mercy, we
are paying a debt of justice.”19
To perform corporal works of mercy, these
general principles apply: (1) monitor job
displacement caused by internal and external
factors; (2) provide reasonable access to
necessary education and training; (3) measure
impact on beneficiaries and on workers, not
merely intention alone; and (3) mitigate biases
and eliminate all forms of unjust discrimination.
Fifth, liberation through structures of grace,
as opposed to structures of sin, forms another
sociological pillar from CST. The Compendium
states,
The moral dimension of the economy
shows that economic efficiency and
the promotion of human development
in solidarity are not two separate or
alternative aims but one indivisible goal.
Morality, which is a necessary part of
economic life, is neither opposed to it
nor neutral: if it is inspired by justice
and solidarity, it represents a factor of
social efficiency within the economy
itself. The production of goods is a duty
to be undertaken in an efficient manner,
otherwise resources are wasted. On the
other hand, it would not be acceptable to
achieve economic growth at the expense
of human beings, entire populations or social groups, condemning them to
indigence. The growth of wealth, seen
in the availability of goods and services,
and the moral demands of an equitable
distribution of these must inspire man
and society as a whole to practise the
essential virtue of solidarity, in order to
combat, in a spirit of justice and charity,
those “structures of sin” where ever they
may be found and which generate and
perpetuate poverty, underdevelopment
and degradation. These structures are built
and strengthened by numerous concrete
acts of human selfishness.20
We might think that the antidote or
prophylaxis to structural sin is liberation
through structures of grace. Thus, Benedict
XVI writes in Caritas in Veritate:
The development of peoples is intimately
linked to the development of individuals.
The human person by nature is actively
involved in his own development. The
development in question is not simply
the result of natural mechanisms, since
as everybody knows, we are all capable
of making free and responsible choices.
Nor is it merely at the mercy of our
caprice, since we all know that we are a
gift, not something self-generated. Our
freedom is profoundly shaped by our
being, and by its limits. No one shapes
his own conscience arbitrarily, but we all
build our own “I” on the basis of a “self”
which is given to us. Not only are other
persons outside our control, but each one
of us is outside his or her own control. A
person's development is compromised,
if he claims to be solely responsible for
producing what he becomes. By analogy, the development of peoples goes awry
if humanity thinks it can re-create itself
through the “wonders” of technology,
just as economic development is exposed
as a destructive sham if it relies on the
“wonders” of finance in order to sustain
unnatural and consumerist growth. In the
face of such Promethean presumption, we
must fortify our love for a freedom that is
not merely arbitrary, but is rendered truly
human by acknowledgment of the good
that underlies it. To this end, man needs
to look inside himself in order to recognize
the fundamental norms of the natural
moral law which God has written on our
hearts."21
General principles for the liberation of
humankind through structures of grace
include the following: (1) engage in inclusive,
human-centered design, (2) apply structural
competency to mitigate the social determinants
of disease, and (3) cooperate appropriately with
others to promote the common good.
BRIDGING CST AND GAI: FORMULATION
OF ETHICAL GUIDELINES
Now, I will attempt to synthesize these insights,
particularly the general principles, as ethical
guidelines for the design, development, and
use of gAI in Catholic health care. While
these guidelines’ specificity will be somewhere
between principles and concrete moral norms
that guide specific behaviors or choices, they
should provide practical influence on gAI in
Catholic health care. Overtime, additional
guidelines derived from CST insights may be
warranted; alternatively, the guidelines below
may require further specification or elaboration
in given circumstances. I have included some guiding questions to prompt further reflection
on these themes, too.
To begin, Catholic health care should engage
human-centered design and inclusivity of
gAI. Algorithms, data sets, and machine
learning applications should reflect our diverse
human tapestry, championing inclusivity and
ensuring marginalized communities are not
sidelined. This could include development
of a sense of humanism and a spirituality
of gAI and the related work. For example,
questions to ask in the design, development,
and deployment of gAI could include (1)
How can the design process actively involve
stakeholders from marginalized communities?
(2) What measures are in place to ensure the
AI system does not perpetuate existing biases?
(3) Does the gAI reflect and enrich integrative
human development as a whole (or does it
compartmentalize and deconstruct in a way
that adversely affects the human experience)?
Next, Catholic health care should aim to
empower its workforce and enable continuous
learning opportunities. As gAI reshapes
work functions and workplaces, continuous
training should ensure the workforce remains
relevant and the connections among workers
strengthened. Minimizing the adverse effects
of disruptive technology is also key to the
adoption and use of gAI applications. This
should include translational skills-building
as well as an emphasis on the humanities in
AI. What training programs are available for
employees to adapt to new AI technologies?
How does the organization plan to maintain
the relevance of human skills in an AI-driven
environment?
Beyond its workforce, Catholic health care should establish collaborative partnerships
and practice community engagement.
Collaboration is key. By forging partnerships
and engaging communities, we ensure gAI
is grounded in real-world needs. Thus, it is
important to ask, what partnerships can be
formed to ensure the AI system meets holistic,
real-world needs and promotes the common
good? Moreover, it’s design, development,
and use should not be siloed; rather, it should
be done in a truly participatory, synodal way.
How is community engagement and feedback
integrated upstream and downstream in
the development and deployment of the AI
application?
To ensure gAI is continuously improved and
maintains ethical integrity, Catholic health care
should enact ethical deployment protocols and
transparent governance structures and processes.
Robust governance structures should oversee
gAI, ensuring ethical considerations are integral.
In addition, advocacy efforts should be aimed
at defending the human person and common
good. What governance structures are in place
to oversee the ethical considerations of AI
deployments? How is transparency maintained
in the AI system’s decision-making processes?
Is it clear who trains the AI and how and on
which data sets?
In addition, Catholic health care should
inclusively develop and collaboratively use
choice architecture and enhance a gAIstakeholder’s autonomy. In an AI-augmented
world, human agency remains paramount.
This entails proper disclosure to appropriate
parties of gAI practices, opt-in versus opt-out
protocols, and informed consent procedures.
Leadership should be able to answer, how are
stakeholders involved in the decision-making process related to AI’s choice architecture? And,
what mechanisms are in place to ensure that an
AI system enhances rather than diminishes user
(or beneficiary) autonomy?
In the rapidly evolving realm of artificial
intelligence, Catholic health care should
clarify proper authenticity and veracity
of AI-generated output. To aid this, the
delineation between authenticity and truth
becomes paramount. Authenticity, in this
context, refers to the genuine origin or source
of data, ensuring that the foundational
elements of AI models are rooted in proper
attribution exhibiting coherence and cogency
(e.g., hallucinations are identified, studied to
understand errors in the AI, and mitigated).
Truth, on the other hand, pertains to the
accuracy and fidelity of AI outputs. As AI
systems increasingly influence decision-making
in healthcare, it is crucial to address and actively
mitigate biases that might skew these outputs.
This not only ensures that the results reflect
genuine realities but also guards against the
inadvertent perpetuation of existing disparities.
Furthermore, a transparent disclosure of
data sources, emphasizing their authenticity
and representativeness, becomes an ethical
imperative, fostering trust and credibility in
AI-driven processes. What protocols are in
place to verify the authenticity of data used and
generated by the AI application? How does the
application ensure that AI-generated output is
accurate and truthful?
Moreover, the reliability of AI systems
transcends their initial accuracy; it encompasses
their consistent performance over time.
Therefore, as these systems become integral to
healthcare, Catholic health care should ensure
continuous monitoring and validation to maintain gAI reliability. Establishing feedback
mechanisms, where users, patients, and other
stakeholders can report inconsistencies or
anomalies, enhances the system's adaptability
and resilience. This iterative process of
validation and recalibration not only ensures
the system's ongoing reliability but also fortifies
trustworthiness. Trust, in this domain, is
not merely about technical robustness; it's
about building and nurturing a relationship of
dependability with communities of concern,
ensuring that they can confidently rely on AI
outputs for critical health care decisions. What
are the key performance indicators for assessing
the reliability of the AI application? How do
these intersect with existing health care related
indicators? When and at what cadence should
the AI application be audited for performance
and compliance with key legal and ethical
norms?
Lastly, as AI delves into realms of creativity
and innovation, the boundaries of intellectual
property and creative rights come to the fore.
Catholic health care should exhibit proper
attribution of AI output as well as choose opensource versus proprietary models in ways that
promote the common good and defend social
justice. Especially in cases where AI models
generate content or make decisions based on
pre-existing works, it becomes ethically and
legally imperative to provide clear attribution
to the original sources. Respecting the creative
rights of individuals and entities ensures that
AI does not inadvertently infringe upon or
dilute the value of original creations. Moreover,
the ethical landscape of AI is further nuanced
by the dichotomy between open-source and
proprietary models. While open-source models
champion transparency and collaborative
betterment, proprietary models underscore the sanctity of intellectual property. Navigating this
landscape requires a delicate balance, ensuring
that the benefits of AI are harnessed without
compromising the rights and contributions
of original creators. What guidelines are in
place for attributing authorship or artistic
credit to the output generated by the AI
system? How does the choice between opensource and proprietary models align with the
organization’s commitment to social justice and
the common good? When would a proprietary
model be justified for the fiscal security of the
organization and under what conditions would
this be effectuated?
CONCLUSION
As we stand at the precipice of another
technological advancement and ethical
discernment, the teachings of the Catholic
Church offer a beacon of light, guiding
our path. The rise of generative AI, with
its transformative potential to ignite an
‘industrialization of thought,’ calls us to
navigate this new frontier with a moral compass
rooted in centuries of wisdom. By grounding
our approach in the pillars of Catholic Social
Teaching, we are better equipped to ensure
that AI serves not just as a tool but as an
extension of our commitment to human
dignity, the dignity of work, the common
good, and the overall betterment of society. In
embracing these ethical guidelines, we affirm
our responsibility to harness the power of AI
in ways that uplift humanity, honor our shared
values, and pave the way for a future where
technology and ethics walk hand in hand.
NICHOLAS KOCKLER, PH.D, MS, HEC-C
Vice President, System Ethics Services
Providence St. Joseph Health
Renton, Washington
[email protected]
ENDNOTES
- Second Vatican Council, Gaudium et spes (GS), Vatican.
va, December 7, 1965, accessed August 30, 2023, no.
4.
- Jiao Sun, et al., “Investigating Explainability of Generative
AI for Code through Scenario-based Design,” IUI ’22,
March 22-25, 2022, DOI: 10.1145/3490099.3511119,
accessed August 31, 2023.
- Sara Vaezy, “Generative AI: Technology Worthy
of Leadership’s Attention,” Boardroom Press,
governanceinstitute.com, August 2023, accessed
October 18, 2023.
- Hazem Zohny, et al., “Ethics of generative AI,” Journal of
Medical Ethics, jme.bmj.com, February 2023, 49 (2): 79,
accessed August 31, 2023.
- Supra note 2.
- Katja Thoring, et al., “The Augmented Designer: A
Research Agenda for Generative AI-Enabled Design,”
International Conference on Engineering Design, July 24-
28, 2023, Cambridge.org, accessed August 31, 2023.
- John Paul II, Centesimus annus, Vatican.va, May 1, 1991,
no. 11, accessed August 31, 2023.
- GS, no. 26.
- John Paul II, Laborem Exercens, Vatican.va, September
14, 1981, no. 9, accessed August 31, 2023.
- Francis, Laudauto Si’ (LS), Vatican.va, May 24, 2015, no.
128, accessed August 31, 2023.
- Second Vatican Council, Dignitatis Humanae, Vatican.va,
December 7, 1965, no. 3, accessed August 31, 2023.
- Lisa Sowle Cahill, Theological Bioethics: Participation,
Justice, Change (Washington, DC: Georgetown University
Press, 2005), 7.
- Pontifical Council for Justice and Peace, Compendium of
the Social Doctrine of the Church, Vatican.va, accessed
August 30, 2023.
- Benedict XVI, Caritas in Veritate (CV), Vatican.va, June
29, 2009, Chapter 4, Paragraph 18, accessed August
31, 2023.
- LS, Chapter 4, Paragraphs 137-139. In paragraph 136,
he writes, “In the same way, when technology disregards the great ethical principles, it ends up considering any
practice whatsoever as licit. As we have seen in this
chapter, a technology severed from ethics will not easily
be able to limit its own power."
- John Paul II, Sollicitudo rei Socialis, Vatican.va,
December 30, 1987, no. 38, accessed August 31, 2023.
- Pius XI, Quadragesimo Anno, Vatican.va, May 15, 1931,
no. 79, accessed August 30, 2023.
- Pontifical Council for Justice and Peace, Compendium
of the Social Doctrine of the Church (Compendium),
Vatican.va, May 26, 2006, no. 182, accessed August 31,
2023.
- Compendium, no. 184.
- Compendium, no. 332.
- CV, Chapter 6, no. 68.