PHIL 123 AI & Ethics

09 / 18

Duty, Consequences, and Moral Choice

Kantian Deontology

Imagine a student named Sam who is applying for a scholarship due Friday. The rough notes describe caring for a younger sibling while working part time, participating in a water-quality research project, and considering a career in environmental science. All three experiences are real, although the notes sound ordinary.

A chatbot turns them into polished prose in seconds. The new draft says Sam “led a research team,” converts family caregiving into a story of “financial hardship” and “resilience,” and presents a career goal with a certainty Sam does not feel. Sam could submit that version, revise it toward what can be defended, or start over.

The first reaction is that Sam should “just be honest.” That slogan leaves the hard part unresolved: one meeting did involve leadership, the family does face financial pressure, and hydrology remains a possible career. Kantian deontology asks which principle would govern the choice and whether that principle could be defended before the people relying on the application.

From Medieval Natural Law to Enlightenment Autonomy

Earlier chapters approached ethics through Aristotle and Thomas Aquinas. Aristotle located moral life in character, habit, practical wisdom, and human flourishing. Aquinas, writing in the thirteenth century, brought Aristotle’s account into a Christian intellectual world. Human beings still have characteristic capacities and goods, but those goods now belong to a created order governed by divine wisdom. In Aquinas’s terms, natural law is the rational creature’s participation in eternal law, grasped through practical reason and used when deciding how to act. Aquinas therefore transforms Aristotle as he adopts him. Virtue and flourishing remain central, while eternal law, creation, sin, grace, and a final end in God supply a wider theological frame.

Immanuel Kant wrote more than five hundred years later, during the European Enlightenment. His short 1784 essay What Is Enlightenment? gives the movement a memorable instruction: “Have courage to use your own reason!” Enlightenment, on his account, requires people to think without handing their understanding over to a book, pastor, physician, or other guardian.

Kant was raised in a Lutheran Pietist household in Konigsberg, Prussia. Pietism was a reform movement within German Lutheranism that emphasized inward devotion, moral seriousness, and self-examination. The Stanford Encyclopedia of Philosophy’s account of Kant’s religion explains both the lasting influence of that setting and the scholarly disagreements about its extent. This background helps explain Kant’s concern with motive, conscience, and the difference between outward conformity and inward moral commitment.

Philosophers usually classify Kant’s mature ethics by its rational grounding and account of autonomy. The Groundwork seeks a principle that binds rational agents without requiring them to accept Lutheran doctrine, Christian revelation, or any church’s authority first. Kant continued to write seriously about God, faith, and Christian teachings. His ethical project nevertheless gives morality priority: reason identifies the moral law, and religious belief must remain compatible with it. His Protestant setting belongs to the history of the view; it does not function as a premise every student must accept before using the argument.

Aquinas and Kant share an ambitious claim. Human reason can recognize moral requirements that bind us even when our immediate preferences point elsewhere. They explain that authority differently. Aquinas places practical reason within a created order of human goods and eternal law. Kant asks what practical reason can require through its own form of lawgiving. His answer is rational autonomy and universal moral law. The moral status of an action depends on its governing principle and whether rational agents could will that principle as universal law.

Autonomy is Kant’s term for rational self-legislation: the will recognizes a principle as binding because it can be willed as universal law. Everyday speech often uses autonomy for personal choice or self-expression. Kant uses it for a will governed by moral law it gives itself through reason.

Heteronomy occurs when desired outcomes, approval, happiness, religious reward, or another incentive supplies the will’s law. A person acting only from fear of punishment may obey the right rule while allowing that fear to govern the choice. In the Groundwork, Kant associates heteronomy with looking for moral authority outside the fitness of one’s maxims for universal law.

The contrast is sharpest between Aquinas and Kant because Aquinas develops rationally knowable human goods within a created order, whereas Kant grounds obligation in the rational will’s capacity for universal law. Even when their conclusions converge, their premises explain moral authority differently. Kantian autonomy requires the agent to recognize the governing principle through reason.

Kant’s central ethical text is the Groundwork of the Metaphysic of Morals (1785). It is freely available in translation through Project Gutenberg and in a plain-language paraphrase from Early Modern Texts. The Stanford Encyclopedia of Philosophy’s entry on Kant’s Moral Philosophy provides a scholarly overview for checking the interpretations used here.

Duty and Good Will

Kant begins the Groundwork with familiar moral distinctions. People often distinguish an honest action done because honesty is right from the same action done only to avoid punishment. They also recognize that intelligence, courage, or wealth can become dangerous when the person using them has a corrupt purpose. Kant does not assume every moral intuition is correct. He starts with ordinary moral judgment and asks which principle could make its claims about good will and duty coherent and universally binding. The Stanford Encyclopedia of Philosophy describes this as an attempt to identify the precise principle underlying common ideas such as good will and duty.

Kant looks to the will because moral evaluation concerns what an agent chooses on principle. In this context, will does not mean a passing wish or a burst of determination. It means the capacity to choose an action according to reasons and practical principles. A rational will can ask whether its reason for acting is one that any rational agent could accept. Kant therefore treats the will as practical reason at work in choice.

In the opening pages of the Groundwork, Kant argues that a good will is the only thing good without qualification (4:393). Intelligence may support harm, courage may serve cruelty, and happiness may accompany something monstrous. A will oriented toward moral principle remains good across those changes in circumstance.

The claim can sound strange because misleading conduct sometimes produces a favorable outcome. A business might profit from fabricated reviews, and its customers might remain satisfied with the product. Kant still asks which principle governed the choice. Respect for moral law gives the action moral worth; a convenient alignment between duty and personal advantage does not.

For Kant, a moral law is a principle that reason recognizes as universally binding on rational agents. It is not a statute passed by a government or a command issued by a school, employer, parent, or church. Those rules may agree with morality, but their institutional source does not make them morally authoritative. A moral law binds even when violating it would be legal, profitable, popular, or difficult to detect.

Kant distinguishes acting in accordance with duty, where conduct matches what morality requires, from acting from duty, where recognition of that requirement governs the choice. In Kant’s specific sense, respect for moral law gives the latter action its moral worth.

Fear of being caught could lead a business to remove fabricated reviews in a way that merely accords with duty. Removing them because buyers are owed truthful testimony would be done from duty. Kant assigns full moral worth only to the choice governed by respect for moral law, even though the outward behavior matches.

Kant makes that distinction so sharp because the right action can coincide with motives whose force depends on the circumstances. Self-interest may support honesty while enforcement is strong. Sympathy may support it while the affected people remain vivid. If a business stops using fabricated reviews only because the fine is expensive, its reason for stopping disappears when enforcement becomes unlikely. The conduct matches duty, but profit and risk still govern the will. In Kant’s technical vocabulary, the action lacks full moral worth because morality itself was not the decisive reason. An action from duty would remain governed by the principle against deception when lying became advantageous and hard to detect. The Stanford Encyclopedia of Philosophy’s account of duty and moral worth describes this as treating duty as a sufficient reason for acting.

Kant’s claim does not require a cold or emotionless moral life. A person may care about buyers and take satisfaction in acting honestly. The disputed question is which reason has final authority when motives conflict. On one influential interpretation, an action expresses good will when the agent would refuse deception because it is wrong, even if feelings or self-interest changed. Critics argue that Kant gives duty too much priority over love and concern. His defenders reply that affection may cooperate with duty as long as the moral requirement remains decisive.

You may recall that Aquinas used the term “law” in his Summa Theologica. It’s important to understand the difference between his idea of natural law vs. Kant’s moral law. Aquinas’s account of natural law begins with God’s created order. Human reason can recognize the goods and purposes built into human nature. It then judges actions by whether they support the flourishing proper to that nature. Virtues help a person choose those goods consistently.

The Groundwork begins elsewhere, with rational agency. Rational beings can step back from impulse and ask which principle should govern their choice. When self-interest or another person’s command supplies the rule, Kant calls the will heteronomous. A moral principle must be one the agent can justify through reason for rational agents generally.

That makes the rational will both the author and the subject of moral law. It is the author because it legislates the principle by willing it as universal law. It is the subject because that universal law binds the lawgiver as fully as it binds everyone else. Self-legislation gives no license to invent convenient private rules. The will can author only principles it could accept as binding on every rational agent. Kant thinks this joint role explains how moral law can obligate a free person. The law arises through the person’s rational agency, while its universal form prevents it from becoming personal preference.

The difference between Kant and Aquinas lies in what reason examines. Aquinas reasons from human goods within creation, while Kant reasons from the form of a maxim and its fitness to become universal law.

Kant directs ethical attention to the principle governing the action. He calls that principle the maxim.

Maxims

A maxim, in Kant’s usage, is the practical principle a person is actually acting on. Every intentional action has one, whether the person could articulate it or not. Kant’s rough template runs like this: in circumstance C, I will do action A for purpose P. When you can spell out the maxim in that form, you can start to test it.

Generating text with a chatbot describes the outward behavior. The maxim must also include the circumstances and purpose, which can make similar uses morally different.

One person might use a chatbot to translate an honest message for a reader. Another might use it to fabricate credentials in order to gain an advantage. Deadline pressure could produce a third maxim: accept whatever the chatbot writes when the risk of being caught seems low.

The generated text may look similar from the outside in each version. Kantian analysis therefore requires an honest account of the maxim the action expresses. A maxim written so carefully that it cannot fail may describe the action a person wishes they were taking.

Kant sharpens this with a distinction between hypothetical and categorical imperatives. A hypothetical imperative is a rational demand that applies when an agent chooses or wills an end. “If you want people to buy the product, you should describe it persuasively.” That command only applies insofar as sales remain the end you are pursuing. A categorical imperative binds you independently of any optional end. “You should not fabricate testimony in order to persuade people” is categorical in structure. It does not disappear when honesty conflicts with what you want.

Kant’s core claim, developed in the Groundwork around 4:414 to 4:416, is that moral obligations are categorical and bind rational agents across changes in preference. A categorical imperative, he writes, “commands a certain conduct immediately, without having as its condition any other purpose to be attained by it.” The command does not depend on first wanting profit, approval, happiness, or some other result. Its authority comes from the principle’s status as a requirement of practical reason. That is why a categorical duty can oppose an agent’s strongest desire and still count as binding.

Three Formulations of the Categorical Imperative

Kant presents several formulations of one moral law. The course uses three of them, summarized below.

A rational agent is a being capable of acting for reasons, forming maxims, setting ends, and recognizing principles that make claims on action. Human beings are imperfect rational agents in Kant’s framework. We can understand a rational requirement while allowing fear, appetite, convenience, or self-interest to govern the choice. Kant’s formulations ask us to test maxims from the standpoint of rational agency, including the agency of people affected by the action.

You will use CI-1 and CI-2 most often in PHIL 123 arguments. CI-3 explains why autonomy is central to Kant. Appetite, fear, convenience, institutional pressure, and commands from others produce heteronomy when they govern the will. An autonomous rational will recognizes moral law as its own.

The Formula of Universal Law

Kant’s Formula of Universal Law directs rational agents to act only on maxims they can also will as universal law (Groundwork 4:421).

Students sometimes reduce Universal Law to the question, “What if everyone did that?” Kant tests whether the maxim can coherently govern rational agents in similar circumstances. Predicting bad results from widespread copying would use consequentialist reasoning.

A simpler case involves AI-generated fake reviews. In August 2024, the U.S. Federal Trade Commission finalized a rule banning fake reviews and testimonials, including AI-generated reviews that misrepresent themselves as coming from real customers. The full rule in the Federal Register and the FTC’s plain-language Q&A cover the rule’s broader scope.

Imagine a small business owner considering exactly that. The maxim would be something like: “When I need sales, I will invent customer testimony so that others will trust my product.” Now apply the Formula of Universal Law. Ask what happens when you treat that maxim as a principle for every rational agent in a similar situation.

The maxim depends on buyers treating customer reviews as reports of real experience. Under a universal law permitting sellers to invent such testimony whenever they need sales, buyers would treat those reviews as seller-written marketing. The maxim would destroy the credibility it needs in order to succeed. Kant classifies this kind of failure as a contradiction in conception.

Duties generated by contradictions in conception are perfect duties, which bind strictly. Imperfect duties also bind but leave more judgment about when and how to fulfill them.

A maxim can fail the Formula of Universal Law in a second way. Sometimes a maxim can be imagined as a universal practice without contradicting itself, but a rational agent still cannot coherently will it. Consider a maxim of never helping anyone else even when you could do so at little cost. You can imagine a world in which nobody ever helps anybody, and it would still exist. But you cannot coherently will such a world, because as a rational agent with ends of your own you sometimes need help, and willing a universal refusal to help is willing against your own rational agency. Kant calls this a contradiction in will. The duties it generates are what he calls imperfect duties. They bind us, but they leave room for judgment about when and how to fulfill them.

For this course, retain the distinction between the two failures. A maxim may destroy its own preconditions, or it may create a world that rational agents could not coherently will.

Back to the Application

Sam’s application mixes accurate claims, stretched claims, and possible inventions. That mixture makes the maxim harder to formulate than it was in the fake-review case.

Begin with Sam’s actual maxim. “When my rough notes underrepresent what I have done, I will use AI to clarify my experience for reviewers” could govern similarly situated applicants without undermining scholarship applications. Reviewers would still receive testimony about actual applicants. The maxim depends on whether each claim still reports the applicant’s experience.

A maxim that permits applicants to make ordinary experiences sound more impressive than they can defend produces a different result. As universal law, it would turn applications into competitive fiction and undermine their role as evidence about applicants. The practice would lose the trust the maxim needs.

Sam’s moral judgment depends on an honest statement of the maxim. A conveniently narrow formulation protects the preferred action from the test and leaves the Kantian argument untested.

The Formula of Humanity

The Formula of Universal Law focuses on the principle. Kant’s second formulation focuses on the persons involved. The Formula of Humanity requires treating humanity in ourselves and others as an end, never merely as a means (Groundwork 4:429).

The word “merely” limits Kant’s prohibition to uses that bypass rational agency. Buying coffee relies on a barista’s work; taking a course relies on a professor; entering a contract relies on another party’s cooperation. Those roles can respect agency because they are disclosed and the people involved retain the ability to consent, set terms, and pursue their own ends within the interaction. Kant therefore allows ordinary instrumental use when it remains compatible with respect for persons.

The moral problem is using someone merely as a means by bypassing that person’s rational agency through deception, coercion, or manipulation. Pauline Kleingeld’s 2020 article on the mere-means prohibition defends one influential actual-consent interpretation: the agent’s principle must make use of another person conditional on that person’s consent. The SEP entry on treating persons as means surveys alternative interpretations. This chapter uses Kleingeld’s account as one way to test respect for rational agency, while recognizing that it is not the only settled reading of Kant.

The fake-review case shifts attention to the buyers’ rational agency. A seller who invents customers treats buyers as a market to be steered. Buyers cannot consent to relying on fabricated testimony because its source has been hidden from them. Their agency has been bypassed. The Formula of Humanity says this makes the action morally suspect regardless of whether the product happens to satisfy them.

The same formulation leaves room for responsible AI use. A seller might use AI to translate an authentic review with the reviewer’s approval, preserving both its source and the buyer’s ability to judge it. The test distinguishes kinds of AI-assisted action without banning the category.

The Formula of Humanity also reaches cases that the Universal Law test handles less well. Consider the labor conditions under which many current AI systems are built. Reporting by Billy Perrigo at TIME on OpenAI’s use of Kenyan contract workers documented workers paid low wages to review disturbing content for training the systems that filter ChatGPT’s outputs. The parties involved have disputed parts of that reporting, so the details deserve careful reading before students draw broad conclusions from that one case. Mary L. Gray and Siddharth Suri describe a wider pattern in Ghost Work, and the Data & Society episode on the book explains that apparent automation can depend on low-paid, often invisible human labor doing tagging, moderation, and evaluation.

The Formula of Universal Law can be applied here, but the result depends on the evidence used to formulate the maxim. “Employ contract workers to label data” describes an activity without naming the conditions or purpose that might make it objectionable. If evidence showed that an employer knowingly concealed psychologically harmful working conditions in order to reduce costs, that fuller maxim could be tested as a universal principle. The Formula of Humanity reaches the documented concern more directly by asking whether workers are treated as agents whose informed consent, safety, and ends matter within the arrangement.

Autonomy, Kingdom of Ends, and Digital Agency

Kant’s account of morality rests on a specific idea of persons as rational agents capable of autonomy: the capacity to give oneself moral law through reason. This differs from doing whatever one wants. An autonomous agent recognizes a moral principle as binding through reason and acts from that recognition.

Heteronomy helps explain a broader danger in AI-assisted judgment. Fear, convenience, institutional pressure, or the system’s confident wording become heteronomous influences when they supply the law for a person’s action. Autonomy requires testing the maxim as a principle that can be owned as universal law. A system may assist with analysis or expression, while the rational agent remains responsible for the moral principle governing the choice.

The Kingdom of Ends gives autonomy a social form. Kant asks us to imagine rational agents as members of a shared moral community, each subject to universal law and each participating in rational lawgiving. The relevant question is whether the maxim behind an AI-assisted decision could belong in a community where the people making and affected by the decision respect one another as rational agents.

Dignity, for Kant, follows from that capacity. Rational agents have what he calls dignity, a kind of value that cannot be priced or exchanged. He develops this idea in the Groundwork around 4:434 to 4:436, and the SEP entry on Dignity tracks how the concept has been picked up in later moral and political philosophy. The rough claim is that persons are not interchangeable with anything else and cannot be treated as if their value were fully captured by what they can do for you.

A chatbot can shape deliberation by supplying polished reasons and strategic framing. Deliberate deception and uncritical delegation are different failures: the first adopts a maxim that cannot be defended, while the second lets convenience or confidence govern the will. Careful revision counts as autonomy only when respect for moral law governs what the person keeps or rejects. The system’s production of an answer does not settle the issue; the maxim governing the person’s use of it does.

Kant gives us a way to analyze rational agency, autonomy, and dignity whenever a digital system helps a person communicate or persuade. The framework keeps responsibility with the rational agent whose maxim governs the action.

Building a Kantian Argument

The OpenStax chapter on deontology gives a general overview of this framework. PHIL 123 uses the compact P1/P2/C form to make an ethical argument inspectable:

That compact form is a starting point. P2 often combines more than one claim. Consider this sentence: “Inventing AI-generated customer reviews treats buyers merely as means.” It sounds descriptive, but it already classifies the action in Kantian terms. The observable facts are that a seller created testimonials from customers who do not exist and presented them as independent reports. The further claim that this arrangement bypasses buyers’ rational agency is a bridge premise. It connects those facts to the Formula of Humanity. A critic can accept the facts while rejecting the bridge.

Framework terms inside a case premise require separate support. Words such as coercive, deceptive, universalizable, and merely as a means do more than report what happened. They interpret the facts through a moral standard. That interpretation may be defensible, but it needs its own reason. A student who writes “the fake reviews manipulate buyers” must identify what information was concealed and explain why that concealment bypasses rational agency. Otherwise, the premise quietly assumes the Kantian judgment that the argument is supposed to establish.

The same problem appears in Universal Law arguments. A student might write:

The argument is compact, but P2 leaves most of the reasoning unstated. It does not state the seller’s maxim, identify the alleged contradiction, or explain which facts make that contradiction relevant. Opening the argument makes those steps visible:

F1 and F2 identify what happened. M1 states the principle governing the seller’s action. B1 explains why that maxim would undermine the practice it needs when treated as universal law, and N1 supplies Kant’s standard. The resulting conclusion targets fabricated testimony without condemning every use of AI in marketing.

A Formula of Humanity argument would organize the reasoning differently. It would retain the case facts, then defend a bridge about the buyers’ agency. The normative standard would prohibit treating persons merely as means. The conclusion would target fabricated testimony while permitting AI assistance that preserves the source of the testimony.

Written in expanded form, the Humanity argument could look like this:

B1 remains open to challenge. A critic might argue that buyers already expect promotional language to be selective and persuasive. The Kantian must then explain why ordinary marketing differs from concealed fabrication. Expanding the argument brings that dispute out of P2 and into view.

The formulations can agree about a case while placing pressure on different parts of the reasoning. Universal Law asks whether the maxim can function as a universal principle. Humanity asks whether the action bypasses rational agency. Autonomy and the Kingdom of Ends help explain who can own the principle and whether rational agents could share it as members of a moral community. Choose the formulation that fits the problem, and make the bridge visible enough for someone else to challenge.

Where the Argument Can Break

Kantian arguments look tidy once the premises have been numbered. Their weak points become visible when a critic asks whether the maxim and bridge were stated fairly. The SEP entry on Kant’s moral philosophy surveys the wider disputes. The fake-review case lets us work through one objection in full.

A critic could accept F1 and F2 while challenging B1. Even if universal fabrication made customer reviews less credible, sellers could still publish them, and some buyers might still believe them. The practice becomes less effective without becoming logically impossible. On this account, the argument has only predicted an undesirable consequence. A contradiction in conception requires an internal conflict in the maxim. The argument risks turning Universal Law into the consequentialist question of whether widespread copying would make the practice work poorly.

A Kantian cannot establish a contradiction merely by predicting lost trust. The reply must show that the maxim both invokes and cancels the convention it needs. The seller presents a review as independent testimony while claiming permission to invent the testimony whenever doing so helps sales. Universalizing that permission removes the distinction between customer testimony and seller-authored promotion that the act asks buyers to rely on. The maxim therefore depends on a public practice that its own rule would erase.

The repair strengthens B1, but it does not make every dispute disappear. A critic could still argue that enough buyers would remain credulous for the deception to succeed. The Kantian response is that occasional success does not rescue a principle that cannot be publicly shared as the rule governing the practice. If that response remains unconvincing, the Formula of Humanity prohibits concealed fabrication because it prevents buyers from evaluating the seller’s appeal on its actual terms and treats their rational agency as an obstacle to be bypassed.

Test that repair against a nearby case. Suppose a business uses AI to translate a verified review, correct its grammar, and secure the reviewer’s approval before publication. With that approval, the translated testimony remains attributable to a real customer who can endorse it. Its maxim preserves the convention that reviews report customer experience. Fabricated testimony invokes that convention while secretly discarding its source. The argument now distinguishes editorial help from fabricated testimony.

This repair revises B1 by explaining the contradiction the maxim creates within the practice it invokes. Other objections may require a different response. A student can defend a challenged premise with evidence, revise the bridge, narrow the conclusion, or concede that part of the original argument fails. The response should change the part of the argument the objection successfully pressures.

Other Kantian pressure points remain. A critic may challenge the framework’s rigidity in cases where duties appear to conflict, dispute whether a person was actually treated merely as a means, or question how a rational-agency-centered framework applies to children, people with cognitive impairments, non-human animals, and AI systems. Each criticism needs a target. “Kant is too rigid” is only a complaint until the critic identifies the duty, case fact, or inference that produces the unacceptable result.

Rigidity becomes a developed objection when the critic supplies a case in which following a duty appears to produce a morally intolerable result. A conflict between keeping a promise and preventing grave harm, for example, pressures the claim that a duty binds without exception. A Kantian reply must show that the maxims or duties were described incorrectly, explain how one duty governs the case, or accept the cost. Repeating that duty is categorical does not answer the objection.

The mere-means objection works differently. It targets the bridge from facts to a claim about rational agency. The critic may agree that another person was used while denying that deception, coercion, or nonconsensual use occurred. Here the dispute turns on what the person knew, what the agent’s principle required, and whether consent was genuine. Kleingeld’s account sharpens that dispute, but rival interpretations can still yield different verdicts. The argument must identify which interpretation it uses.

What You Retain for Yourself

In the discussion and midterm ahead, you will build Kantian arguments about your own dilemmas and cross-examine arguments made by classmates. A chatbot may help generate candidate maxims, bridges, or objections. Each output remains a proposal until you test it against the chapter and your case.

A fluent objection may attack a maxim you are not using. It may confuse an unfavorable consequence with a contradiction in the maxim. It may call every use of another person a mere-means violation. Before accepting the output, identify its move. Does it challenge a premise, challenge the inference, or offer a counterexample? Then point to the exact line in your argument that receives the pressure.

The chatbot does not know whether your maxim honestly represents your reason for acting. It cannot settle whether a counterexample is close enough to matter or whether the relevant people retain meaningful agency. Those judgments require your knowledge of the dilemma and your understanding of Kant. The output can make a possible weakness visible. You decide whether the weakness is genuine and what the argument should become.

For the midterm, write the compact argument first. Open it if P2 hides a maxim, bridge, prediction, or moral classification. Ask the chatbot for a strong objection, then verify that objection against the Kantian concepts in this chapter. Decide whether to defend, revise, narrow, or concede part of the argument. Record what you changed and why. If the objection misstates your maxim or substitutes consequence-counting for Kantian reasoning, reject it and explain the error.

Keep a short record of the exchange. Note the argument line the chatbot challenged, the chapter concept used to evaluate that challenge, and the revision you accepted or rejected. This record makes your judgment visible. It also lets you notice whether the chatbot helped expose a real weakness or merely supplied polished philosophical vocabulary.

Kantian deontology gives you one way to make a call. State the principle behind the action. Test whether it can hold as universal law. Ask whether persons are respected as ends. Make the bridge between the facts and the Kantian judgment visible. Then decide what remains defensible after a serious objection. The final judgment, and responsibility for the argument supporting it, remains yours.

References

Primary and Supporting Texts

  • Kant, Immanuel. 1785. Fundamental Principles of the Metaphysic of Morals. Translated by Thomas Kingsmill Abbott. Project Gutenberg. Primary-source translation used to cross-check the chapter’s account of Kant; this edition does not supply Akademie pagination.
  • Kant, Immanuel. 1785. Groundwork for the Metaphysic of Morals. Adapted and prepared by Jonathan Bennett. Early Modern Texts, 2005–2008. Plain-language paraphrase used only as an accessibility aid.
  • Kant, Immanuel. 1784. “What Is Enlightenment?”. Translated by Lewis White Beck, Robert E. Anchor, and Emil L. Fackenheim. German History in Documents and Images. Primary text used for the chapter’s explanation of Kant’s Enlightenment setting and the call to use one’s own reason.
  • Johnson, Robert, and Adam Cureton. 2025. “Kant’s Moral Philosophy.” Stanford Encyclopedia of Philosophy. Scholarly overview used to check the chapter’s account of duty, the Categorical Imperative, autonomy, the unity of Kant’s formulations, and the standard Akademie locators cited in the chapter.
  • Pasternack, Lawrence, and Courtney Fugate. 2025. “Kant’s Philosophy of Religion.” Stanford Encyclopedia of Philosophy. Scholarly overview used to qualify the relationship among Kant’s Lutheran Pietist upbringing, mature moral philosophy, and continuing treatment of religion.
  • Kleingeld, Pauline. 2020. “How to Use Someone ‘Merely as a Means.’” Kantian Review 25 (3): 389–414. Primary scholarly source for the agent-focused actual-consent interpretation discussed in the chapter.
  • Kerstein, Samuel. 2023. “Treating Persons as Means.” Stanford Encyclopedia of Philosophy. Scholarly survey used to place Kleingeld’s interpretation within the wider debate.
  • OpenStax. 2022. “Deontology.” In Introduction to Philosophy. OER overview used as supplemental reading.

Background Sources

Applied Cases

Scholarly layer

Person record