UPSC Mains 2026 General Studies Paper IV — Model Answers

The ethical core. A doctoral examiner is appointed in a personal capacity because the university needs a named human judgment on the originality and quality of the work. That judgment is non-delegable. The professor's wrong is not that he used a tool; it is that he substituted the tool for the judgment he was appointed to exercise, and did not disclose it.

From the perspective of accountability

• The chain of answerability is broken. If the report is wrong, unfair or defamatory, no one can be asked why a particular conclusion was reached. The professor did not reach it; the system cannot explain it, and cannot be sanctioned.
• Some modifications is not authorship. Editing an output the examiner cannot independently justify is ratification, not evaluation. Accountability requires that he be able to defend every claim in the report on his own reading of the thesis.
• Non-disclosure removes the institution's ability to act. Had the university known, it could have set conditions, sought a second examiner or refused. Concealment transfers risk to the candidate and the institution without their knowledge.
• Data and confidentiality. An unpublished thesis uploaded to an external system may breach the confidentiality obligations of examinership and the candidate's rights over unpublished work.

From the perspective of integrity

• Role morality. Integrity is coherence between the role accepted and the conduct performed. Accepting the fee, the honour and the authority of examinership while outsourcing its substance is a false representation of work as one's own, and Kant's test is decisive here: the professor is treating the candidate as a means to the discharge of an inconvenient obligation.
• Paucity of time is a reason, not a justification. The ethical response to an unmanageable workload is to decline or seek an extension, not to fabricate diligence.
• The double standard. A scholar penalised for undisclosed AI use in the thesis is being judged by an examiner who did the same, which corrodes the moral authority of the entire assessment system.
• Harm to a specific person. A doctoral report shapes a career. An unverified evaluation, hallucinated citations, a misread argument, a confident but wrong assessment of novelty, falls on one identifiable candidate.

The defensible rule. The answer is not prohibition. AI can legitimately assist with language, formatting, literature mapping and plagiarism screening. The line is that evaluative judgment must be human, informed by actually reading the work, and disclosed where assistance was used. Universities should require an explicit declaration of AI assistance by examiners as well as candidates, verifiable audit of examiner engagement, and workload norms that make honest examination possible. Integrity is easier to sustain when the system does not quietly reward the shortcut.

As District Magistrate I would face conflicts of duty, not a simple choice between right and wrong.

Development against dispossession. The project may bring investment, revenue and employment to a poor district. The same project removes a community from the habitat that constitutes its livelihood, its social structure and its cosmology. Utilitarian aggregation, greater good for greater numbers, is the argument that will be pressed on me, and its weakness is precisely that it permits a concentrated, irreversible loss to a small community to be outweighed by diffuse gains to others. Rawls's test, that an arrangement must be justifiable to those worst affected, is the corrective I would apply.

Consent against consultation. The Forest Rights Act, 2006 requires recognition of individual and community forest rights and Gram Sabha consent for diversion of forest land, and PESA, 1996 requires consultation before acquisition in Scheduled Areas. The pressure in practice is to convert consent into a formality, to hold the Gram Sabha in a manner that produces the desired resolution. Presiding over a procedurally valid but substantively hollow consent would be the most likely and most serious ethical failure available to me.

Duty to the state against duty to the citizen. Political and administrative superiors may want speed; the affected community wants voice. My legal duty is to the Constitution rather than to the government of the day, but the professional cost of that distinction is borne personally.

Compensation against restitution. Land can be valued; a forest-based way of life cannot. Cash compensation to people with no experience of handling capital frequently produces impoverishment rather than security. The honest question is not the rate of compensation but whether livelihood can actually be restored.

Conflict of interest and inducement. A large business house is a source of pressure, of information asymmetry and, sometimes, of inducement to me and to my subordinates.

Public order. If protest follows, I hold the coercive power of the state. Using it against people resisting their own displacement, while the law is on their side, is the sharpest dilemma of all.

Truthfulness. I will be expected to give the community assurances about jobs and rehabilitation that I do not control and may not be able to keep.

How I would hold the line. Constitutional morality over expediency; full and honest disclosure of the project to the community in its own language and before decisions harden; strict compliance with FRA and PESA in substance, not form; independent social impact assessment; rehabilitation designed as livelihood restoration and not as a payment; recording my objections in writing where I am overruled; and dealing with protest through negotiation, with force as a last resort governed by proportionality. Where I cannot in conscience implement a lawful order, the honourable course is to record dissent and seek transfer, not to sabotage quietly.

Position. No. The District Magistrate should not authorise the use of this research on the tribe in the manner proposed. He may, and should, do a great deal else.

Under medical ethics

• Autonomy and informed consent. Consent must be individual, informed, voluntary and revocable. An endangered tribe, possibly a particularly vulnerable tribal group, faces language barriers, unfamiliarity with biomedical concepts and an authority gradient so steep that a District Magistrate's suggestion is experienced as an instruction. Consent obtained in those conditions is unlikely to be genuinely voluntary, and community assent through a headman does not substitute for individual consent.
• Non-maleficence. Primum non nocere. Clinical trials are not yet conducted, so the supplement's dose, toxicity, interactions and long-term effects are unknown. A mineral supplement is not automatically benign; minerals have narrow therapeutic windows and real toxicity at excess.
• Beneficence. The expected benefit is real but hypothetical. Beneficence requires a favourable balance of expected benefit over expected harm, and that balance cannot be computed before trials.
• Justice. The most serious concern. Using an isolated, dependent population to generate evidence that will benefit others is the historical pattern of exploitative research, from Tuskegee onwards. The Nuremberg Code, the Declaration of Helsinki and India's ICMR National Ethical Guidelines for Biomedical and Health Research, 2017 all require that vulnerable populations be protected from bearing research risk disproportionately, and that research on a community be relevant to that community's own health needs and confer benefit on it.
• Regulatory reality. Any interventional use requires Ethics Committee approval, registration on the Clinical Trials Registry-India and, where applicable, approval under the New Drugs and Clinical Trials Rules, 2019. A District Magistrate has no authority to bypass this, and none to consent on another person's behalf.

Under administrative ethics

• Role and competence. The DM's duties are protective and facilitative. He is not a research sponsor, and he is not competent to assess a trial protocol. Acting outside role is itself an ethical failure, however good the motive.
• Coercion by office. State power carries implicit compulsion. Even a well-meant recommendation from the District Magistrate is not, for this community, a free choice.
• Paternalism and dignity. Deciding for a community because it is thought incapable of deciding for itself is the colonial posture the Constitution's tribal safeguards were written against.
• Trust as a public asset. If harm followed, the loss of trust would extend to immunisation, nutrition and every other health programme in the district for years.

What the DM should do instead

1. Ensure immediate, proven care: nutritional assessment, diagnosis of the deformity, screening for fluorosis, vitamin D deficiency, genetic and environmental causes, and treatment of what is already treatable.
2. Investigate environmental causation, water quality and diet, which for skeletal deformity in an isolated population is often where the answer lies, and act on it.
3. Refer the research to ICMR and the state health department so that, if warranted, a properly designed and ethically cleared study is conducted with community engagement, translated consent, independent monitoring and assured post-trial access.
4. Guarantee that the community will receive the benefit if the remedy is proved, and is not merely the site where it is tested.
5. Document and be transparent with the community throughout, in its own language, so that the choice remains theirs.

Conclusion. The urgency is real and it argues for faster proper research, not for skipping it. Compassion that bypasses consent is not compassion; it is a well-intentioned form of the same disregard that endangered the tribe in the first place.

Start by rejecting the zero-sum framing. National security is not a rival value to human rights; it is the condition in which rights can be exercised. A state that tortures, disappears or detains without trial in the name of security destroys the very thing it claims to defend, and it manufactures the grievance that sustains the threat. Equally, a state that cannot protect life leaves the most basic right unsecured. The question is therefore not which value prevails but what limits a democracy accepts on itself while acting against genuine threats.

The reconciling principle: proportionality. Any restriction on rights in the name of security should satisfy four tests, the same structure the Supreme Court applied in Puttaswamy (2017):

• Legality. The action rests on a law, publicly known and precise, not on executive discretion.
• Legitimate aim. The purpose is genuine security, not political convenience, and the burden of showing this lies on the state.
• Necessity. No less restrictive measure would achieve the aim.
• Proportionality in the narrow sense. The harm imposed is not disproportionate to the benefit gained, and the restriction is temporary, reviewable and reversible.

Institutional mechanisms that make the principle operative

• Judicial oversight. Habeas corpus, judicial authorisation for surveillance and detention, and reasoned orders. ADM Jabalpur (1976) is India's cautionary precedent for what happens when courts abdicate, and its overruling in Puttaswamy is the correction.
• Legislative scrutiny. Sunset clauses on emergency powers, parliamentary committee oversight of intelligence agencies, and mandatory reporting of the use of exceptional powers.
• Independent human rights institutions. The NHRC and State Commissions, with real access to security establishments, and enforceable guidelines of the kind laid down in D. K. Basu v. State of West Bengal (1997) on arrest and custody.
• Accountability of forces. Training in human rights, functioning internal grievance systems, and prosecution of violations without the shield of blanket immunity. Impunity is what converts an isolated abuse into a recruitment argument for the adversary.
• Transparency to the extent possible. Publication of aggregate data on detentions, encounters and interceptions, so that secrecy protects operations rather than errors.
• Redress. Compensation and rehabilitation for wrongful detention and custodial harm, which is what distinguishes a state that errs from a state that does not care.

In the Indian context. The Constitution itself contemplates the balance: Article 22 gives procedural safeguards on arrest while permitting preventive detention within limits; Articles 19(2) to 19(6) allow reasonable restrictions in the interest of sovereignty and security; Article 21 sets the floor. The unresolved questions are concrete and worth naming: the scope of preventive detention laws and of the UAPA's bail provisions, the operation of AFSPA in disturbed areas and the Justice Jeevan Reddy Committee's recommendation on it, and surveillance without statutory authorisation, which the Digital Personal Data Protection Act, 2023 has not settled.

The practical wisdom. Security measures should be targeted rather than communal, temporary rather than permanent, reviewable rather than final, and accompanied by remedy. Where a whole community is treated as suspect, the state loses the intelligence cooperation on which counter-terrorism actually depends. The ethical position and the effective position converge, which is the strongest argument for holding it.

Ambedkar's argument. In his closing speech to the Constituent Assembly on 25 November 1949, Ambedkar said that with the coming of the Constitution "we must abandon the method of civil disobedience, non-cooperation and satyagraha", because "where constitutional methods are open, there can be no justification for these unconstitutional methods", and that these are "nothing but the Grammar of Anarchy". His reasoning is not authoritarian. It is that in a democracy the majority's will has a lawful channel, that extra-constitutional pressure by any group invites the same by every other, and that a young democracy cannot survive competitive coercion.

The counter-position. Civil disobedience has a serious moral pedigree. Thoreau grounded it in conscience, Gandhi in satyagraha as truth-force with self-suffering rather than injury to the opponent, and King in the distinction between a just and an unjust law. Rawls gave it the most usable definition for a democracy: a public, non-violent, conscientious act contrary to law, undertaken to bring about a change in law or policy, by persons who accept the legitimacy of the constitutional order and are willing to bear the legal consequences. On this account civil disobedience is not a rejection of the constitution but an appeal to its own principles when ordinary channels have failed a minority.

Where the tension actually lies. Ambedkar's premise is that constitutional methods are open. They may be formally open and substantively closed: a permanent minority outvoted on every question, a legislature that does not debate, a judiciary too slow or too costly for the affected, an executive that ignores representation. In such conditions insisting on constitutional methods alone means asking the excluded to accept exclusion indefinitely.

How democracies should navigate it: criteria for legitimacy

1. Exhaustion of ordinary means. Petition, litigation, representation and electoral contest should have been genuinely tried, not gestured at.
2. Non-violence, strictly. Violence forfeits the moral claim entirely and converts a movement into coercion. This is where Ambedkar's warning bites hardest.
3. Publicity and openness. Civil disobedience is done openly, not covertly; concealment marks it as evasion rather than appeal.
4. Acceptance of legal consequences. Willingness to be arrested and punished is what demonstrates fidelity to the legal order and distinguishes the disobedient from the criminal.
5. Proportionality and specificity. The action targets the objectionable law or policy, and does not paralyse unrelated public life or take third parties hostage.
6. A justice claim, not an interest claim. The demand appeals to constitutional principles, equality, dignity, rights, rather than to sectional advantage.
7. Openness to dialogue. A movement that refuses all negotiation has abandoned persuasion for compulsion.

And the corresponding duties of the state. A democracy earns Ambedkar's argument by keeping constitutional channels genuinely open: consultation before legislation, functioning committees, accessible courts, and negotiation rather than force as the first response to protest. It must also police the difference between disobedience and destruction, and it must not treat every protest as a security threat, which is the mirror-image failure.

Conclusion. Ambedkar was right that the habit of extra-constitutional pressure is corrosive, and Rawls is right that principled disobedience can be a corrective within constitutional loyalty rather than a rejection of it. The two are reconciled by the criteria above: civil disobedience is legitimate where it is a last resort, non-violent, public, accepts punishment and appeals to constitutional principle. Anything short of that, and Ambedkar's phrase applies.

The doctrine. Gandhi's trusteeship holds that whatever a person possesses beyond need, wealth, talent, or position, is held in trust for society rather than owned absolutely. He proposed it as a moral alternative to both unrestrained capitalism and expropriation: not the abolition of the possessor but the transformation of possession into stewardship. Applied to public office, the doctrine reads directly: an officer owns nothing of what he administers. Authority, budget, information and discretion are entrusted, and they must be used for the beneficiary, not the trustee.

Ways an administrative officer can apply it

• Office as trust, not property. A posting is not a personal entitlement, a fiefdom or a reward. This rules out treating a district or a department as territory, resisting oversight, or regarding scrutiny as an insult.
• Public resources as fiduciary. Every rupee, vehicle, hour of staff time and square metre of public land is held for a beneficiary who is not present in the room. The test of any expenditure is whether the trustee could justify it to that beneficiary, which is a sharper standard than mere legality.
• Gandhi's talisman as a decision rule. Recall the face of the poorest and weakest person you have seen, and ask whether the step you contemplate will be of any use to him. This converts trusteeship into a usable test at the moment of decision, particularly in prioritising among competing claims.
• Discretion exercised for the absent. Where the law leaves room, the trustee resolves in favour of the person with least capacity to influence the outcome, not the one who is present, articulate and well-connected.
• Transparency as accounting. A trustee renders accounts. Proactive disclosure under Section 4 of the Right to Information Act, published beneficiary lists, social audit and public hearings are the administrative form of that duty.
• Non-possession and simplicity. Aparigraha in practice means declaring assets, refusing gifts and hospitality that create obligation, and modesty in the use of official facilities. It also means not accumulating power for its own sake.
• Means and ends inseparable. Gandhi's most demanding rule for an administrator: a good outcome achieved by suppressing dissent, falsifying a record or coercing consent is not a good outcome. Fairness lies in the process as much as the result.
• Succession and institution-building. A trustee prepares for handover. Leaving systems, documentation and trained staff rather than a personal legend is the trusteeship test applied to a tenure.
• Trustee for the future. Environmental and fiscal decisions bind people not yet born, who are the ultimate absent beneficiaries.

The standard criticism, and the answer. Trusteeship is often dismissed as naive, dependent on a change of heart that cannot be relied upon, which is why Nehru preferred structural regulation. The criticism is fair against trusteeship as an economic system. It is much weaker against trusteeship as an ethic of office, because public administration already imposes fiduciary duty in law: the AIS Conduct Rules, the Prevention of Corruption Act and the constitutional oath all assume that authority is held for others. Gandhi supplies the internal disposition that makes those external rules self-enforcing, which matters because no rulebook can anticipate every use of discretion.

Conclusion. Trusteeship does not tell an officer what the right decision is. It tells him whose interest the decision must serve, and it removes the illusion that authority is personal. That is precisely what fairness in governance requires.

The line and its speaker. In Henry IV, Part 1, Act V Scene 2, Hotspur tells his followers on the eve of Shrewsbury, "Now, for our consciences, the arms are fair, / When the intent of bearing them is just." It is a commander steadying men before battle, and Shakespeare's placement is deliberately ironic: Hotspur is a rebel, he is certain of his justice, and he is dead by the end of the play. The dramatist offers the maxim and its refutation together.

The ethical claim. The line asserts that the justice of the intent legitimises the instrument. It is a familiar structure: the just war tradition's requirement of right intention, the state's monopoly on legitimate force, and the everyday administrative belief that a hard measure is defensible if the purpose is public good.

Where it holds in governance. Coercive power is intrinsic to the state, and cannot be renounced. Police force, taxation, land acquisition, quarantine, preventive detention, prosecution: each is an instrument that would be a wrong if used for a private purpose. Intent is genuinely part of what makes them legitimate. An officer who orders a lathi charge to protect life is not doing the same act as one who orders it to punish dissent, even if the physical act is identical. Motive matters, and mens rea in criminal law, and the doctrine of malice in administrative law, both encode that.

Where it fails, and why the failure is dangerous

• Sincerity is not justification. Everyone who wields power believes their intent is just; Hotspur did. A test that relies on the actor's own assessment of his motives is not a test at all.
• Means shape ends. Gandhi's objection is the deepest: means and ends are not separable, "the means may be likened to a seed, the end to a tree". Surveillance built for terrorism is used for dissent; emergency powers taken for a crisis outlive it. The instrument reshapes the institution that uses it.
• Kant's objection. Persons may not be used merely as means, however good the purpose. A just intent does not license torture, fabricated evidence or the sacrifice of an innocent for public benefit.
• The historical record. The most serious abuses of state power in modern history were carried out by people convinced of the justice of their cause. Sincerity is not a safeguard; it is often the mechanism.

The implications for governance. The maxim is acceptable only when hedged, and the hedges are the answer to the question:

1. Legality. Intent never substitutes for authority. The instrument must be one the law permits, exercised by the person the law authorises.
2. Proportionality and necessity. The least coercive effective means, and a harm not disproportionate to the good sought.
3. Objective justification. Reasons recorded in writing and testable by someone else, so that justice of intent is demonstrated rather than asserted.
4. Independent review. Judicial and legislative oversight, precisely because self-assessment of motive is unreliable.
5. Absolute limits. Some means remain impermissible whatever the intent: custodial torture, fabricated evidence, collective punishment.
6. Temporariness and reversibility. Exceptional instruments should carry sunset clauses, because the intent that justified them will not be inherited by everyone who later holds the power.

Conclusion. Shakespeare's line describes a psychological truth about how power justifies itself to itself, and Shakespeare knew it. In governance, just intent is a necessary condition for the use of coercive means and never a sufficient one. What makes an instrument fair is intent plus legality, proportionality and accountability, and the last three exist precisely because the first cannot be verified from inside.

Beneficence. One of the four principles of biomedical ethics set out by Beauchamp and Childress, beneficence is the positive duty to act for the patient's benefit, to prevent and remove harm and to weigh benefits against risks. It is more than the negative duty of non-maleficence: the physician must actively promote welfare, not merely avoid injury.

Paternalism. Paternalism is the overriding of a person's known preferences or choices, by another, for that person's own good. It is soft where the person's decision-making is impaired, by unconsciousness, incapacity or ignorance, and intervention protects a choice they cannot presently make. It is hard where the person is competent and informed, and their choice is overridden anyway. Medical ethics accepts soft paternalism in narrow circumstances and rejects hard paternalism, because it treats a competent adult as unable to run her own life.

Analysing the doctor's action

• The beneficent element is genuine. The doctor knows the serious adverse effects of long-term medication. Disclosing them fully and recommending surgery is exactly what beneficence, combined with the duty of informed consent, requires. A doctor who withheld that information to avoid influencing the patient would be failing her.
• The decisive move is the route taken. The patient had made a choice: medication. Rather than persuading her, the doctor persuaded her family, who then brought pressure to bear. That converts the family from a support system into an instrument of compliance, and it is the ethically decisive act in the case.
• Reluctant consent is a warning sign, not a resolution. Consent extracted through social pressure from those a patient depends on is not voluntary in the sense autonomy requires. Formally the consent form is signed; substantively the choice was not hers. In India this pattern has a gendered edge: a woman's medical decisions are routinely mediated through husband and in-laws, and a doctor who works through that structure is amplifying an existing inequality rather than protecting a patient.
• Which paternalism is this? It is hard paternalism conducted indirectly. The patient was competent; her preference was known; it was overridden by engineering the social environment around her. The indirection makes it worse, not better, because it is harder to identify and to resist.

The verdict. The doctor's motive is beneficent and her clinical judgment may well be correct. Her method violates autonomy, which the four-principle framework treats as the presumptive trump in a competent adult, and it does so in a way designed to avoid confronting the patient's refusal directly. Well-intentioned and ethically wrong are not mutually exclusive, and this case is the standard illustration.

What she should have done instead

1. Full disclosure to the patient herself: risks of long-term medication, risks and benefits of surgery, alternatives, and the consequences of doing nothing, in language she understands.
2. Time and a second opinion, since a decision made under acute pressure is rarely a free one.
3. Family involvement only with the patient's consent, and as support rather than as leverage.
4. Explore why she prefers medication, which is usually fear of surgery, cost, dependants at home or past experience, and address the real obstacle.
5. If she still refuses, document informed refusal and continue to treat her. A competent adult's right to refuse recommended treatment is settled, and honouring it is not abandonment.

Conclusion. Beneficence tells the doctor what to recommend; autonomy determines who decides. When beneficence is pursued by manipulating the decision-maker's environment rather than by persuading the decision-maker, it has become paternalism, and the patient's sense of betrayal is the reliable symptom.

The distinction, and its order of priority. Peter Drucker's formulation is the one the question uses: efficiency is doing things right, a ratio of output to input; effectiveness is doing the right thing, the fit between action and purpose. The two are not equal partners. Effectiveness is logically prior, because efficiency in pursuit of the wrong objective merely reaches the wrong place faster. Drucker's own gloss stands: there is nothing so useless as doing efficiently that which should not be done at all.

How each fails without the other

• Efficiency without effectiveness. A department clears files within the prescribed time and reports full compliance while the citizen's problem remains unsolved. Toilets built and counted, but not connected to water. Classrooms constructed and enrolment achieved, while children cannot read. This is the characteristic Indian administrative failure: the input and output indicators are excellent and the outcome is poor.
• Effectiveness without efficiency. The objective is correct but the cost is unsustainable. Relief that reaches everyone at ten times the necessary expenditure crowds out the next programme. In a resource-constrained state, waste is not merely untidy; it is a decision to deny something else.

Striking the balance

1. Sequence, do not average. First establish the purpose, then optimise the method. Ask what outcome is sought, for whom, and what would count as success, before asking how to do it faster or cheaper.
2. Measure outcomes, not just outputs. Learning levels rather than enrolment; functional water supply rather than pipelines laid; disposal with redress rather than files closed. What is measured is what gets managed, so the choice of indicator determines which value dominates.
3. Let context set the weight. In an emergency, disaster response, epidemic, law and order, speed is effectiveness and efficiency should dominate. In policy design, project appraisal and rights adjudication, effectiveness must dominate and haste is the enemy.
4. Periodic re-examination. Schemes acquire momentum. A standing review that asks whether the objective is still the right one prevents efficient pursuit of an obsolete purpose.
5. Use technology for the efficiency half and judgement for the effectiveness half. Digitisation, workflow automation and single-window systems raise throughput; they do not tell you what to do. Reserve human deliberation for the question of purpose.
6. Feedback from the beneficiary. The person the service exists for is the only reliable judge of whether the right thing was done. Grievance data, social audit and citizen feedback close the loop that internal reporting cannot.
7. Align incentives. If appraisal rewards expenditure and file disposal, officers will optimise for those. Outcome-linked appraisal is what makes effectiveness rational for the individual officer and not just for the system.

Productivity, properly understood. Productivity is not maximum output per unit of input; it is maximum public value per unit of input. That definition contains both terms: value keeps effectiveness in view, per unit of input keeps efficiency in view. An administrator who holds that definition will not be tempted to trade one for the other.

Conclusion. Effectiveness sets the direction and efficiency sets the pace. Direction first, then pace, and the discipline to keep re-asking whether the direction is still right.

Why both extremes fail. Enforce strictly and most schools in the remote area close; the children do not transfer to better schools, because there are none within reach. They drop out, and girls drop out first. Do not enforce, and the state signals that standards are optional, poor children are educated in unsafe buildings by unqualified teachers, and the discrimination becomes permanent. A rule enforced into closure and a rule abandoned both deny the same child the same right.

The reframing that resolves it. Regulations under the Right of Children to Free and Compulsory Education Act, 2009 exist to guarantee quality, not to certify buildings. Article 21A makes free and compulsory education for children aged six to fourteen a fundamental right. When the literal enforcement of a subordinate rule defeats the fundamental right the rule was made to serve, the administrator's duty runs to the right. That is not discretion to ignore the law; it is fidelity to its purpose, and it is why administrative law recognises the difference between the letter and the object of a provision.

How I would proceed

1. Grade the norms by what they protect. They are not equivalent.
   • Non-negotiable, immediately: structural safety of the building, safe drinking water, functioning toilets particularly for girls, protection from abuse, and mid-day meal hygiene. These bear directly on life and dignity, and no closure-avoidance argument justifies waiving them.
   • Negotiable with a time-bound plan: playground area, boundary wall, land extent, prescribed room dimensions, library and laboratory norms. These are quality goals achievable on a schedule.
   • Requiring support, not punishment: teacher qualification and pupil-teacher ratio, which cannot be met by closing a school but can be met by deputation, distance training and recruitment incentives for remote postings.
2. Assess before acting. A school-by-school audit rather than a blanket order, identifying which norm is breached and why. In remote areas the cause is usually absence of state provision, not defiance.
3. Issue conditional recognition with a compliance calendar and monitored milestones, rather than closure notices.
4. Fix the supply side. Use Samagra Shiksha funds, MPLADS and MLALADS, CSR and district mineral funds for buildings, water and toilets; deputation and residential incentives for teachers.
5. Escalate honestly. Where the norm itself is unrealistic for the terrain, report upward with evidence and seek a reasoned relaxation for remote and tribal areas, which is a policy decision, not a field-level favour.
6. Never close without an alternative. No school ceases to operate unless the children have a functioning, reachable place to study, with transport or residential provision arranged.
7. Be transparent. Publish the compliance status of every school. Parents are entitled to know the standard their children are being taught in, and publication creates pressure that inspection alone does not.

The ethical principles at work. Purposive rather than literal interpretation; proportionality, so that the response matches the gravity of the breach; the best interests of the child as the paramount consideration; equity, since applying an identical standard to unequal circumstances entrenches inequality; and accountability, in that relaxation is documented, time-bound and reviewable rather than informal.

Conclusion. The administrator's task is not to choose between the rule and the child. It is to enforce immediately what protects the child, to schedule what improves the school, and to fix what the state itself failed to provide. Discretion exercised transparently, on record and against a deadline is administration; the same discretion exercised quietly and indefinitely is dereliction.

Why the claim is largely accurate. The international system has no sovereign above states, so each is ultimately responsible for its own survival. Governments are accountable to their own citizens, not to humanity, and a leader who sacrifices national interest for a universal principle will be replaced by one who does not. Add the security dilemma, in which defensive measures appear threatening to others, and the result is that strategic interest systematically outranks moral norms. The record supports the description: selective invocation of human rights depending on whether the violator is an ally, arms sales to regimes whose conduct is condemned in the same week, climate commitments diluted at the point of implementation, vaccine hoarding during the pandemic, and Security Council vetoes that protect friends from accountability.

The ethical positions

• Realism, from Thucydides and Kautilya to Morgenthau, holds that the statesman's supreme moral duty is the security and welfare of his own people; applying private morality to statecraft is a category error, and Machiavelli's argument is that a ruler who is good among the wicked ensures his own ruin. On this view the subjugation of ethics to interest is not a lapse but a duty.
• Cosmopolitanism, from Kant's Perpetual Peace to Singer and Pogge, holds that moral obligation does not stop at borders. Human rights are universal, and treating foreigners as morally discountable is arbitrary, as arbitrary as discounting people by caste or race.
• The middle position, sometimes called the English School or ethical realism, holds that states are self-interested but operate in an international society with shared rules, that reputation is a real asset, and that long-run interest and ethical conduct converge more often than short-run interest suggests.

Where ethics does constrain states, and why

• Reciprocity. International humanitarian law, diplomatic immunity and treaty compliance are honoured largely because violation invites violation. Self-interest can produce ethical behaviour without ethical motive.
• Reputation. A state that breaks agreements finds fewer partners. Credibility is a strategic asset, which is why even powerful states argue that their actions are lawful rather than admitting they are not.
• Domestic politics. Citizens, courts, media and civil society constrain foreign policy from within; the moral cost of an alliance is paid at home.
• Institutions. The UN, ICC, WTO dispute settlement and human rights machinery are weak but not empty; they raise the cost of misconduct and give the aggrieved a forum.
• Long-horizon problems. Climate, pandemics, oceans and antimicrobial resistance cannot be solved by any state alone, and here narrow interest is simply irrational.

The ethical conclusion. The right response to the statement is neither cynicism nor moralism. Cynicism, treating the description as permission, is self-defeating, because a state that is known to honour nothing cannot build coalitions and cannot rely on others honouring anything. Moralism, demanding that states act as saints, is a policy for irrelevance and cannot be sustained by any elected government. The defensible position is enlightened self-interest with a floor: pursue national interest, but accept limits that are non-negotiable, no complicity in genocide, no torture, compliance with humanitarian law, honouring of agreements, and consistency in applying standards to friend and adversary alike. Consistency is the practical test of whether a state has ethics or only interests, because it costs something.

The Indian case. India's tradition holds both strands honestly. Ashoka's dhamma vijaya, Kautilya's mandala theory, Gandhi's insistence on means, Nehru's Panchsheel and non-alignment, and the contemporary framing of Vasudhaiva Kutumbakam and Voice of the Global South all sit alongside a hard-headed defence of strategic autonomy, a refusal to sign the NPT, and continued energy purchases through sanctions. India's position is precisely the middle one: ethics as a long-term interest and as a source of legitimacy, not as a substitute for capability.

The case for empathy, stated fairly. A child with a diagnosed learning disability depends on continuity of instruction from a teacher who understands her needs, and replacing that teacher can set her back by months. The Rights of Persons with Disabilities Act, 2016 requires reasonable accommodation and inclusive education, and specific learning disability is among the twenty-one recognised conditions. A rigid transfer policy that ignores such consequences is itself a moral failure. Empathy, the capacity to enter another's situation, is a foundational value for civil service precisely because rules cannot anticipate every case.

What is actually wrong here. The problem is not that the DM felt empathy. It is the route by which the case reached him and the interest of the person who brought it.

• Conflict of interest. The request came from the DM's Personal Assistant, a subordinate in a relationship of daily dependence and personal proximity. The PA is not a neutral petitioner; he is a direct private beneficiary, since the teacher gives private tuition to his daughter.
• Private tuition is itself likely a breach. Most state service rules prohibit government teachers from taking private tuition. The administration is therefore protecting a teacher in order to preserve an arrangement that may itself be irregular, which compounds rather than mitigates the problem.
• Unequal access to discretion. Other children in the district with learning disabilities, whose parents do not sit outside the Collector's room, have no route to the same relief. The DM has not exercised compassion; he has distributed compassion by proximity, which is the operational definition of favouritism.
• Effect on the transfer system. Transfers are a rationing mechanism for scarce teachers. An exception granted informally at the top invites every other exception to be pursued the same way, and the officer who refuses the next one will appear arbitrary.
• Precedent and integrity of office. The action is invisible to audit, unrecorded and unreviewable, which is the characteristic signature of an ethically compromised decision even when the outcome is benign.

The test that separates empathy from favouritism. Ask three questions:

1. Would I do this for a stranger with identical facts who had no access to me? If not, it is favouritism.
2. Can I record the reason in writing and defend it publicly? If not, it is favouritism.
3. Am I applying a rule to a class of cases, or making an exception for a person? Empathy generalises; favouritism does not.

The DM's decision fails all three.

What he should have done

1. Recuse himself from the request as made, and tell the PA that a personal request through the office is not the appropriate channel.
2. Ask the affected parent to apply through the education department, with the disability certificate and the school's assessment, so the case enters the record.
3. Have the education officer examine it against policy, and if the policy permits deferment of transfer on educational-continuity grounds, apply it, with reasons recorded.
4. Address the class, not the individual. If continuity for children with disabilities is a genuine problem, issue a general instruction: transfers of teachers handling identified children with special needs to be deferred to the end of the academic session, applicable district-wide.
5. Deal with the private tuition separately, on its own merits and under the service rules, rather than allowing it to shelter behind the child's need.
6. Ensure the child's entitlement is met properly through the resource teacher and inclusive education support that Samagra Shiksha provides, so that her education does not depend on one individual or on her father's access to the Collector.

Conclusion. Empathy and compliance are not opposed. The officer's duty is to convert a felt sympathy into an impartial rule, applied to everyone in the same situation, recorded and reviewable. Compassion exercised through personal access is not a higher morality than rule-following; it is a lower one, because it is available only to those who are already close to power.

The distinction. A duty is obligatory: it can be demanded of you, others have a corresponding claim, and failing it is blameworthy. Charity is supererogatory, from supererogare, to pay beyond what is owed: it is good to do, praiseworthy when done, but no one is wronged if you do not. On the conventional view, giving to famine relief is charity; not stealing is a duty. The moral philosophers referred to in the question, Peter Singer above all in Famine, Affluence, and Morality (1972), argue that this classification is mistaken for a large class of cases.

The argument, set out

1. Suffering and death from lack of food, shelter and medical care are bad.
2. If it is in our power to prevent something bad from happening without thereby sacrificing anything of comparable moral importance, we ought, morally, to do it.
3. Money spent on luxuries, beyond what maintains our financial stability, is not of comparable moral importance to a life.
4. Therefore we ought to give such money to prevent that suffering. It is a duty, not an optional kindness.

The intuition pump. Singer's drowning child: you pass a shallow pond in which a child is drowning. Wading in will ruin your expensive shoes. Nobody thinks you may walk on, and nobody thinks that rescuing the child is a fine act of generosity rather than a plain obligation. The distance between you and the child, and the fact that others could also help, do not appear to change the obligation. If they do not, then geographical distance and the presence of other potential donors cannot convert a duty into charity in the famine case either.

Why the reclassification matters. The distinction is not verbal. If it is charity, the giver is praised for giving and not blamed for withholding, and the recipient is a beneficiary of goodwill. If it is duty, withholding is a wrong, the recipient has a claim, and the giver's generosity is not a favour but a payment of what is owed. Kant's language captures the shift: an imperfect duty of beneficence, owed to no particular person and discharged at one's discretion, becomes far more demanding once the means are known and the cost is low. The dignity of the person in need is also restored, since a claimant is not a supplicant.

Objections, and answers

• Demandingness. The principle seems to require giving until further giving would cost something of comparable importance, which is close to impoverishing oneself. Singer offers a weaker version, and the question's own framing supplies it: without significantly affecting your financial stability. That threshold makes the duty demanding but liveable.
• Special obligations. We owe more to our own children and dependants. Granted, and the qualification about financial stability accommodates it. Special obligations limit the duty; they do not abolish it.
• Ineffectiveness and aid dependency. Some aid does harm. That is an argument for giving effectively, to interventions with evidence behind them, not for not giving.
• It is the state's job. True, and individuals should support just institutions. But the duty is not discharged by pointing at an institution that is not in fact discharging it.
• The collective action problem. If everyone gave, each would owe little. The fact that others do not give does not release you, any more than a crowd of bystanders at the pond releases the one person who could wade in.

The Indian parallel. The tradition already contains this distinction, and often collapses it in the same direction. Dana is enjoined as a duty in the Dharmashastra, not left to sentiment. The Gita's nishkama karma prescribes right action performed without attachment to reward, which is duty rather than benevolence. Gandhi's talisman, and his trusteeship doctrine that surplus wealth is held for society, treat sharing as an obligation of the possessor and not as a virtue of the generous. Buddhist and Sikh practice institutionalise it through dana and daswandh and langar, forms in which giving is a structured requirement rather than an impulse.

Conclusion. The claim is justified. Once the cost to me is small and the benefit to another is survival, the moral difference between refusing to save a child in front of me and refusing to save one far away rests on proximity and visibility, and neither is a morally relevant fact. What we call charity in such cases is more accurately described as a duty we have been permitted to treat as optional.

(a) The ethical issues

1. Consent given for one procedure does not extend to another. Lata consented to a diagnostic laparoscopy under general anaesthesia. A hysterectomy is not a larger version of that procedure; it is a different, irreversible operation that removes an organ and ends fertility. The consent taken was real but it was consent to something else.

2. Surrogate consent from a competent adult's relative has no standing. Sujatha is Lata's sister-in-law. Lata is an adult of sound mind, temporarily unconscious because of anaesthesia the doctor herself administered. Indian law recognises substituted consent only for minors and for persons unable to consent, and unconsciousness under anaesthesia during an elective procedure is not that situation, because the patient will be competent again within hours. This case is very close on its facts to Samira Kohli v. Dr. Prabha Manchanda (2008), in which the Supreme Court held that consent to a diagnostic laparoscopy did not authorise a hysterectomy, and that consent obtained from the patient's mother while she was under anaesthesia was not valid.

3. The emergency exception does not apply. A doctor may act without consent where delay would endanger life. A tumour that could be malignant, on a closer examination during surgery, is not that. The alternative Dr. Mansi rejected, taking a biopsy and waiting for the result, was clinically available and safe. Convenience, sparing a second surgery, cost and pain, is a benefit to be weighed by the patient, not for her.

4. Autonomy and bodily integrity. The right to decide what is done to one's body is protected under Article 21 and is the core of the doctor-patient relationship. Removing a uterus without the woman's knowledge is a violation of dignity, not merely a procedural lapse.

5. Beneficence pursued at the cost of autonomy. Dr. Mansi's motive was protective. That makes this hard paternalism, and good motive does not convert an unconsented operation into a consented one.

6. A gendered dimension. Decisions about a woman's reproductive organs were made by a doctor and an in-law while she was unconscious. Reproductive autonomy is exactly where Indian women's medical decisions are most often mediated by family, and a professional who works through that pattern reinforces it.

7. Institutional trust and the response of the police. The police attempted to persuade Lata that good intention settled the matter. It does not, and treating a rights violation as a misunderstanding compounds the original wrong and pushes the citizen towards litigation as the only remedy, which is what happened.

8. Anticipatory duty. The possibility of finding a tumour was foreseeable. A competent clinician should have discussed the contingency before surgery and obtained Lata's instructions on it.

(b) The moral conduct of the doctor

Clinically defensible; ethically indefensible. These are separate findings and both should be stated.

• What was reasonable. Suspecting malignancy, weighing the burden of a second surgery, and wanting to protect the patient from a second exposure to anaesthesia are all sound clinical reasoning. Dr. Mansi acted in good faith, and there is no suggestion of negligence in the surgery itself or of a financial motive.
• Where she failed. She treated a decision that belonged to Lata as a decision that belonged to her. Under time pressure she chose the course that was easiest to justify medically and hardest to justify ethically. Taking Sujatha's signature was not a safeguard; it was the appearance of a safeguard, and obtaining it in writing suggests she knew consent was needed and settled for consent from the wrong person.
• What she should have done. Take a biopsy, complete the laparoscopy, allow Lata to recover, disclose the findings fully, and let her decide on a hysterectomy with time, information and, if she wished, a second opinion. If she consented, operate then. That course loses a few days and preserves everything that matters.
• After the event. The right response to Lata's distress was full disclosure, an unreserved apology, an explanation of the reasoning, and support including counselling, not a defence mounted through the police.

Verdict. Dr. Mansi's conduct was well-intentioned and ethically wrong. She violated informed consent and bodily autonomy in a situation that permitted her to obtain both. Lata's sense of betrayal is not a misunderstanding to be managed; it is an accurate perception of what happened.

What the case should change. Pre-operative discussion of foreseeable intra-operative findings, with the patient's instructions recorded in advance; institutional protocols that prohibit extending surgery beyond consent except in a genuine life-threatening emergency; a documented second-clinician check before any irreversible extension; and training that treats consent as a continuing conversation rather than a signature obtained once.

(a) Ethical issues, including biases, in AI-based data-driven policing

The biases, by mechanism

• Historical data bias. The system learns from past police records. If a neighbourhood was policed more intensively, it generated more arrests, and the algorithm reads that as more crime rather than as more policing. The memorandum's core allegation, that the system rests on incorrect historical data produced by discriminatory policing, is technically the correct criticism.
• The feedback loop. Predicted hotspots receive more patrolling, which generates more recorded incidents, which confirms the prediction. The model becomes self-fulfilling, and its apparent accuracy is manufactured by its own deployment.
• Proxy discrimination. Even with no field for religion, caste or migrant status, postal code, occupation and income act as proxies. Removing the protected attribute does not remove the discrimination.
• Automation bias. Officers defer to the machine because it appears objective, so a contestable inference acquires the authority of a fact and human judgement is suspended precisely where it is most needed.
• Group attribution. Flagging an area means every resident of it is treated as a heightened risk. Suspicion attaches to address rather than conduct.

The other ethical issues

• Presumption of innocence and preventive detention. Acting against people for what a model predicts they may do inverts the basic principle of criminal justice. Preventive detention is a serious deprivation of liberty and it should not rest on a probabilistic score.
• Consent and privacy. Biometric data is being captured from crowds without knowledge or consent. Puttaswamy (2017) requires legality, legitimate aim, necessity and proportionality for any such intrusion, and mass facial and biometric capture in public is very hard to justify as necessary.
• Absence of a legal basis. There appears to be no statute authorising the collection, retention or matching. Deployment on administrative sanction alone fails the legality limb before the others are reached.
• Opacity and lack of due process. Residents do not know what data is held against them, cannot see it, cannot contest it and cannot have it corrected. There is no notice, no hearing, no appeal.
• Accountability gap. If the system is wrong, who is answerable, the vendor, the SP, the constable who acted? Unallocated responsibility is the characteristic ethical failure of algorithmic administration.
• Function creep and data security. A database built for crime detection can be repurposed, leaked or sold, and biometric data cannot be reissued once compromised.
• The utilitarian trap. Public order has visibly improved. That is a real good, and it is exactly the argument that makes rights erosion attractive. Aggregate benefit distributed across a city, purchased by concentrated harm to one poor migrant neighbourhood, is the distributive question utilitarianism cannot answer.
• Erosion of police legitimacy. A community that believes it is under suspicion by algorithm stops reporting crime and stops giving information, and community cooperation is the single most productive source of intelligence any police force has.

(b) In Ravi's position: options and the ethical course

Option 1: Continue as is. Merits: order has improved, no immediate disruption, keeps political and departmental support. Demerits: institutionalises discrimination, is legally vulnerable, deepens alienation, and the improvement may be an artefact of displacement rather than prevention. Rejected.

Option 2: Discontinue the system entirely. Merits: removes the rights violation immediately, answers the memorandum, restores trust. Demerits: discards a genuinely useful capability, risks a deterioration in public order for which Ravi is accountable, and is an overcorrection that would be reversed by his successor. Rejected.

Option 3: Suspend the coercive uses, audit the system, and restart under safeguards. Merits: stops the harm at once without discarding capability; treats the allegation as a factual claim to be tested rather than a complaint to be managed; is defensible in law and in court. Demerits: requires time, expertise and the acceptance of a temporary loss of an operational tool.

Option 4: Refer the whole matter upward and await instructions. Merits: procedurally safe for Ravi. Demerits: the harm continues while the file moves. This is the abdication that looks like prudence.

Recommended course: Option 3, with a definite sequence

1. Immediately suspend the coercive applications — preventive detentions and checkpost siting driven by algorithmic output — while retaining the system in an advisory, non-actioned mode. This stops the deprivation of liberty first.
2. Receive the memorandum formally, acknowledge it in writing, and meet the community leaders, lawyers and activists again with a stated timeline. Treating them as complainants to be pacified is what turned this into a confrontation.
3. Commission an independent audit by a mixed team — a data scientist, a legal expert, a civil society representative and a senior officer from outside the district — to test the training data for historical bias, measure differential error rates across communities, and examine the legal basis for biometric collection.
4. Establish transparency and redress. Publish what categories of data are held and for how long; create a mechanism for a resident to know what is recorded against them, contest it and have it corrected or erased.
5. Rule out prohibited uses permanently. No preventive detention, arrest or use of force on algorithmic prediction alone. Every action must rest on independent, human, articulable grounds recorded in writing.
6. Restore the human in the loop. Officers must be required to record their own reasons, and to be able to justify an action without reference to the system.
7. Report up the chain and seek legal cover. Place the findings before the DGP and the state government and press for a statutory framework, since a lasting solution is not within a Superintendent of Police's power.
8. Rebuild in the affected neighbourhood through community policing, beat officers who are known by name, recruitment from the community, and grievance meetings, so that improvement in order is achieved with the residents rather than against them.

Justification. This course satisfies legality, since it removes actions that lack statutory basis; proportionality, since surveillance is retained only where necessary and the most intrusive uses are stopped; justice as fairness, since the arrangement is now defensible to the group worst affected; and accountability, since human officers remain answerable for every coercive act. It is also the most effective option, because policing that a community regards as arbitrary loses the cooperation that makes policing work. Crucially, it does not treat technology as the villain: it insists that an instrument of state coercion be governed by the same standards as every other, which is what an officer's oath to the Constitution actually requires.

(a) The ethical issues

• Corruption as a denial of rights. A change of land use is an entitlement to be decided on merit. Selling it converts a public function into a private market and imposes an illegal tax on citizens, most heavily on those least able to pay.
• Violation of natural justice and equality. Out-of-turn approvals breach the elementary requirement that like cases be treated alike, and Article 14 is not a slogan here but the standard being violated.
• Harassment through delay. The backlog is not an accident; it is the instrument of extortion. Delay creates the desperation that makes a bribe seem rational. Inaction is the corrupt act.
• Capture of the office by intermediaries. Touts operating through subordinate officials means the state's authority has been privatised at the counter, and the citizen no longer deals with the government at all.
• A climate of fear inside the organisation. Honest employees are afraid to process files. That is the most corrosive fact in the case, because it means the office has lost the capacity to function honestly even where individuals wish to.
• Economic and social harm. Blocked land use holds back local development, and public dissatisfaction erodes the legitimacy of the administration as such.
• Seema's personal dilemmas. Duty to act against duty to keep services running; the risk of institutional resistance and union backlash; the temptation to prioritise a clean personal record over a difficult intervention; and the danger of injustice to subordinates who may be coerced rather than complicit.
• The moral cost of inaction. Ignoring it is not neutrality. It ratifies the arrangement, signals to the network that it is safe, and continues the harassment of citizens. There is no option here that is not a choice.

(b) Options open to Seema

Option 1: Ignore it and focus on other priorities. Merits: no confrontation, administration runs undisturbed, personally safe. Demerits: abdication of the core duty of her office, harm continues and deepens, and her reputation for integrity becomes hollow. Rejected outright.

Option 2: Immediate aggressive raid-and-suspend action. Merits: decisive, visible, satisfies the demand for justice, deters others. Demerits: the case warns of exactly this — institutional resistance, union backlash, mass absence of staff, complete stoppage of an already blocked service, and the likely collapse of hastily gathered cases for want of evidence. Punishing the visible small players while the network reconstitutes is the usual result.

Option 3: Refer the whole matter to the state anti-corruption bureau and await the outcome. Merits: uses specialist investigators, keeps Seema at arm's length from a hostile process. Demerits: slow, uncertain, leaks are likely, and the backlog and the harassment continue meanwhile. Necessary at some stage; insufficient as the first step.

Option 4 (recommended): Reform the process first, build evidence quietly, then act on individuals. Attack the conditions that make the corruption possible before attacking the people, so that removing individuals does not simply create vacancies for their successors.

The recommended course, in sequence

1. Remove the discretion that is being sold. Order strict chronological disposal of applications with a published, real-time online queue showing the position of every file. Where a file is taken out of turn, require written reasons on record. Transparency alone destroys most of the rent.
2. Cut out the tout interface. A single-window receipt counter with computerised acknowledgement and SMS tracking, so no applicant needs to meet a dealing official at all. Bring the service under the state's Right to Public Services Act, with a statutory time limit and penalty for default.
3. Clear the backlog visibly and immediately, through a time-bound camp with officers drawn from other branches. This does two things at once: it relieves the citizen and it removes the scarcity the network profits from.
4. Protect and re-empower the honest majority. Meet the staff, state plainly that lawful processing will be backed by her personally, redeploy the subordinates identified as central to the racket to non-public postings pending inquiry, and give the rest a clean process to work in.
5. Build the case properly. Quietly commission a file audit for the pattern of out-of-turn approvals, secure the records, and refer the evidence to the anti-corruption bureau or vigilance for a professional investigation. Do not attempt a sting or a parallel investigation herself.
6. Deal with the touts as a law-and-order matter, denying them physical access to the Collectorate through the SP.
7. Communicate. Publish the new process, hold an open grievance day, and use local media so that applicants know they need not pay. Public knowledge is what makes the reform irreversible.
8. Handle the union properly. Engage the association in advance, make clear that action is against identified individuals and not against the staff, and that the objective is a workplace where honest work is possible. Most employees benefit from that framing and will not defend the racket.
9. Record everything upward. Keep the Divisional Commissioner and the state government informed in writing, which protects both the reform and Seema.

Justification. This sequence satisfies the constraint the case sets: daily administration is not disrupted, in fact it improves first, so public support builds before confrontation begins. It follows the 2nd ARC's diagnosis that corruption is best attacked by reducing discretion, increasing transparency and ensuring accountability rather than by punishment alone. It is fair to individuals, since action follows evidence rather than suspicion. And it is durable, because a chronological, online, time-bound process leaves nothing to sell after the individuals are gone. Aggressive investigation without process reform removes the players; process reform without investigation removes the incentive but not the impunity. Both are required, and the order matters.

(a) The options before Vikas, and the balance

The real trade-off is between two kinds of error. A system can be tuned to avoid exclusion error, denying a genuine beneficiary, or inclusion error, admitting an ineligible one. Biometric authentication was introduced to cut inclusion error, and it has raised exclusion error, which falls on manual labourers with worn fingerprints, on growing children, on the elderly, and on everyone during a connectivity failure. The two errors are not morally equivalent. An inclusion error costs the exchequer the price of a ration; an exclusion error costs a poor household its food, and the National Food Security Act, 2013 makes that entitlement a statutory right, with Article 21 behind it.

Option 1: Enforce biometric authentication strictly. Merits: leakage minimised, fiscal probity assured, rules followed, Vikas is personally safe. Demerits: genuine beneficiaries denied food through no fault of theirs; likely violation of the NFSA and of Article 21; distress in the poorest households.

Option 2: Leave it to distributor discretion. Merits: the recipient who is known locally gets the grain; humane in the individual case. Demerits: unauditable, invites diversion and favouritism, and makes the citizen's entitlement depend on a shopkeeper's goodwill. Rejected.

Option 3 (recommended): Keep biometric as the default, and build an auditable, layered fallback so no genuine beneficiary is turned away. This is the course that serves both values instead of trading one against the other.

How it works in practice

1. A cascade of authentication, not a single gate. Fingerprint, then iris where available, then a registered mobile OTP, then Aadhaar-linked face authentication. Most failures resolve at the second or third step.
2. Offline and exception mode for genuine failures. Where all electronic modes fail, permit distribution against a recorded exception: entry in a register, the beneficiary's signature or thumb impression, a photograph, and a receipt, with the record synchronised when connectivity returns and audited monthly. Discretion is thus structured and reviewable rather than personal and invisible.
3. Cap and monitor. Publish exception rates by fair price shop. A shop with an unusually high exception rate is flagged for inspection, which is how the fallback is prevented from becoming the new leakage.
4. Fix the causes. Re-enrol and update biometrics for those repeatedly failing, particularly children and manual workers; provide offline-capable ePoS devices with local caching; use alternative connectivity where WiFi is weak; deploy mobile enrolment camps in remote habitations.
5. Nominee and doorstep provision for the aged, the disabled and the bedridden, so that authentication does not become a physical barrier.
6. Grievance and publicity. A helpline and a district grievance officer with a fixed response time, and a notice at every shop, in the local language, stating that no beneficiary may be denied ration solely for biometric failure, with the number to call.
7. Report upward and seek formal cover. Vikas should place the problem, the data and his proposed protocol before the state food department in writing and seek a standing order, so that the fallback is policy and not an individual officer's improvisation.

How to balance probity with empathy. Fiscal probity is served by auditability, not by rigidity. A recorded, monitored exception is auditable; a denial is not more honest than a documented issue. The governing principle is that the burden of a system failure must fall on the state, not on the beneficiary, since the state chose the technology and the citizen did not. Empathy here is not a departure from the rules; it is fidelity to the statute the rules exist to implement.

(b) The ethics of technology as a gatekeeper for welfare schemes

What technology legitimately does. Digitisation has removed ghost beneficiaries, cut intermediaries, enabled portability through One Nation One Ration Card so that a migrant can draw rations away from home, and created records that make diversion harder. These are real gains for the poor and should be stated first, because the critique is not an argument for returning to paper.

What changes ethically when a machine stands at the gate

• The error moves, and so does its incidence. Ghost beneficiaries were a loss to the state. Authentication failure is a loss to the individual, and it falls hardest on those whose bodies bear the marks of manual labour and age. The system's costs are now paid by the poorest.
• Convenience is not consent. Making a technology mandatory for a subsistence entitlement is coercive by construction, however voluntary it is in form.
• The right becomes conditional on infrastructure. An entitlement that depends on connectivity is not an entitlement in a place without connectivity. Constitutionally, the state's obligation cannot be suspended by its own supply failure.
• Accountability becomes diffuse. When the machine says no, no official has said no, and there is no one to appeal to. This is the deepest ethical problem: technology can remove the human against whom a citizen has recourse.
• Dignity. Being told repeatedly that one's fingerprints do not exist, in front of a queue, is a specific indignity, and it deters people from returning.
• Function creep and privacy. Biometric and consumption data collected for one purpose is available for others, and the poor have least capacity to object.
• Efficiency displacing purpose. The metric becomes authentication rate rather than food actually delivered, and administration optimises the metric.

The principles that should govern it

1. Technology may verify a right; it may never extinguish one. No entitlement should be denied solely because a device failed.
2. A non-digital fallback is mandatory in every scheme, designed at the outset rather than improvised at the counter.
3. Publish exclusion data. Authentication failure rates should be reported like any other administrative statistic; what is not measured is not fixed.
4. Human override with accountability, so that discretion exists, is recorded and is auditable.
5. Design for the hardest case first. A system built for the manual labourer's worn fingerprint and the village with two bars of signal will work for everyone else.
6. Consent, minimisation and purpose limitation under the Digital Personal Data Protection Act, 2023, particularly for biometric data.
7. Remember whose convenience it is. Digital welfare architecture is often designed for the administrator's assurance rather than the beneficiary's ease, and the ethical test is whose burden it reduces.

Conclusion. The purpose of the public distribution system is that a poor household eats. Technology is a means to that end and acquires no independent authority. The moment the instrument begins to determine who is entitled, rather than to verify who is, the ethical relationship has been inverted.

(a) The ethical issues to be addressed

• The primacy of life. Deaths have occurred and relief camps are running. The first ethical obligation is not reconciliation but the protection of life, and food, water, medicine and safety in the camps for both communities equally.
• Impartiality when the machinery itself is partisan. The rivalry has entered the local administration and police. An officer whose instruments are compromised cannot deliver impartial administration merely by intending it, and neutrality that is not visible to both sides is worthless in a conflict.
• Perceived versus actual fairness. In communal situations perception is the operative reality. Every act, whose camp is visited first, whose leaders are met, which arrests are made, will be read as a signal. Rajan must be fair and must be seen by both Jhara and Biru to be fair.
• Force and proportionality. Restoring order may require coercion. Its use must be proportionate, minimal, non-discriminatory and documented, and it must not be applied more heavily to one community, which is the commonest failure in Indian communal policing.
• Justice against expedience. Peace bought by shielding perpetrators of one community is not peace; impunity guarantees the next round. Equally, prosecutions perceived as one-sided will reignite the conflict. Even-handed accountability is the only stable position.
• The root causes are ethical claims. Conflicting land claims, uneven resource allocation and inadequate political representation are grievances of justice. Treating the conflict as purely a law-and-order problem addresses the symptom and guarantees recurrence.
• The rights of the displaced. Relief camps must not become permanent segregation. Return, rehabilitation and restoration of livelihood are obligations, not concessions.
• National security. A hostile neighbour may exploit the disorder, which raises the cost of delay and constrains how long a purely persuasive approach can be tried.
• Truthfulness and rumour. In such situations misinformation kills. Rajan has a duty to establish an authoritative, credible information channel, and never to issue a statement he cannot stand behind.
• Personal integrity under pressure. He will face political pressure to favour one side, and the temptation to define success as quiet rather than as justice.

(b) The options before Rajan

Option 1: Immediate large-scale force, curfew and mass preventive arrests. Merits: rapid restoration of order, deterrence, denies the neighbouring country an opening. Demerits: uses a compromised local police whose partiality will show in whom it arrests; risks casualties; converts a communal conflict into a state-versus-people conflict; and is the most likely course to destroy his non-partisan standing on day one.

Option 2: Negotiate first and defer coercive measures. Merits: consent-based, addresses causes, builds durable peace. Demerits: too slow for a volatile situation; leaves the highway closed and supplies blocked; hardliners may read restraint as weakness and escalate.

Option 3: Do the minimum and refer everything to the state government. Merits: avoids personal risk. Demerits: abdication of the responsibility for which he was posted; the situation deteriorates while the file moves.

Option 4 (recommended): A graded, humanitarian-first approach with neutral force and simultaneous dialogue. This is close to the approach the case says Rajan already favours, and it is the ethically strongest.

The recommended sequence

1. Open the highway first, framed as a humanitarian act. Food and medicines for both communities is a purpose neither side can oppose, and it re-establishes the administration's authority on a neutral basis rather than a coercive one.
2. Requisition central forces for neutral tasks. Bring in CRPF and deploy it on the highway, at camp perimeters and along the fault lines. Central forces have no local kinship and are therefore perceived as neutral, which is precisely what the local police cannot be at present. This is the single most important measure for protecting his non-partisan image.
3. Restore neutrality within the machinery. Immediately move officials and police personnel who are demonstrably partisan to non-sensitive duties, in equal measure from both communities, and place mixed teams under officers brought in from outside the district. Symmetry is essential; a one-sided transfer list would itself become an incident.
4. Make impartiality visible and symmetrical. Visit both relief camps the same day, meet delegations of both communities together where possible and separately where not, publish relief distribution figures for both camps, and give identical undertakings to each.
5. Begin talks immediately and in parallel, through a peace committee with equal representation, including women and respected non-political figures from both communities, with a written, published agenda covering land claims, resource allocation and representation.
6. Enforce the law even-handedly and transparently. Register cases and act on evidence, publishing arrest figures community-wise so that even-handedness can be verified rather than merely asserted. Where possible, use a special investigation team from outside the district.
7. Counter rumour actively. A daily official briefing, a control room, monitoring of social media, and prompt, verified correction of false claims.
8. Guard the external dimension through coordination with intelligence agencies and border forces, while keeping this separate from the civil administration's public posture, so that the community is not treated as a security suspect.
9. Address the causes in a time-bound way. A land-records survey and a transparent claims process, an audit of resource allocation between the communities, and support for representation in local bodies, since the case names these as the stated causes.
10. Record and report every significant decision, with reasons, and keep the state government informed, which protects both the process and the officer.

Why this option best protects a non-partisan image without compromising ethics. It leads with an act, opening the highway, whose beneficiaries are both communities, so his first visible decision cannot be coded as favouring either. It uses neutral force rather than compromised local force, which is the difference between restoring order and taking a side. It applies symmetry to every discretionary act, transfers, visits, prosecutions, so impartiality is demonstrable rather than declared. It refuses the false choice between order and justice by pursuing both at once. And it treats the underlying grievances as legitimate, which is what distinguishes a peace that holds from a lull that does not. Neutrality here is not passivity; it is equal firmness towards both, which is the only form of neutrality a district magistrate can honestly offer.

(a) The options, and the choice

The two proposals compared

The criteria that should govern the decision. Arms transfers are not ordinary exports. Ajit's own charter recognises that due diligence at every stage must ensure conformity to national policy and international treaties. The relevant tests are: risk of diversion to non-state actors; risk of use in violation of international humanitarian law; regional destabilisation, including nuclear escalation; end-use verification capacity; consistency with India's declared positions; and the long-term strategic and reputational consequence. These are the same criteria that international practice, including the Arms Trade Treaty's risk assessment framework, uses, and India's own policy of case-by-case scrutiny reflects them.

Applying the criteria to Country B

• Support for guerrilla forces abroad is the decisive fact. It creates a serious and foreseeable risk that Indian long-range artillery and missiles reach non-state armed groups, either by transfer or by capture. For a country that has itself suffered from cross-border support to armed groups, and that presses this case at the FATF and the United Nations, supplying such a state would be a direct contradiction of its own position.
• Non-signatory to the NPT with nuclear weapons and delivery systems. Supplying long-range delivery-capable systems into such an arsenal raises escalation and proliferation risk.
• A military budget that eats into human resources and infrastructure means the sale would be financed at the cost of that country's own citizens' development, which engages an ethical objection independent of security.
• Alliance with a superpower means end-use control and technology security cannot be assured, and sensitive design information may not remain confined.
• The advance payment and the smaller order are commercially attractive and ethically irrelevant.

Applying the criteria to Country A

• Legitimate, articulated defensive need against a specific threat.
• No indication of support to armed groups, no proliferation red flag, no alliance complicating end-use control.
• Non-aligned, which fits India's own tradition and reduces the risk of the transfer being read as taking sides in a bloc rivalry.
• R&D collaboration on next-generation weapons is a durable strategic asset far exceeding the value of one order.
• The long-term loan is a genuine commercial risk, and should be treated as one.

Recommendation: sell to Country A, subject to conditions.

• A binding end-use and end-user certificate, with no re-transfer without prior written consent and a right of on-site verification.
• Credit risk management through sovereign guarantee, export credit insurance and staged deliveries linked to payment milestones.
• Formal inter-ministerial clearance from External Affairs, Defence and Finance, with the file recording the reasons for preferring A over B.
• A written, reasoned communication to Country B, declining on policy grounds applicable to all buyers, so the refusal is principled rather than personal, and offering non-lethal or dual-use cooperation where policy permits.

The strongest counter-argument, answered. It was pressed on Ajit that if India refuses, Country B will simply buy elsewhere, so India forgoes the revenue without preventing the harm. This argument proves too much: it would justify any transaction to any buyer, and it is the standard reasoning by which arms control collapses. Three answers hold. First, complicity is not neutralised by substitutability; a supplier remains responsible for its own act. Second, the harm is not identical, since Indian systems in those hands create an Indian liability, an Indian reputational cost and an Indian traceability problem. Third, and practically, a country known to apply consistent criteria becomes a more attractive long-term supplier to responsible buyers, because they can rely on it not to arm their adversaries either. Reliability is the asset being built.

(b) Balancing economic and strategic interest with ethical considerations

1. Set the criteria in advance and publish them. Ethics applied case by case under commercial pressure collapses. A published policy framework covering diversion risk, humanitarian law, regional stability and end-use makes each decision defensible and removes it from negotiation.
2. Treat ethics as a long-horizon interest, not a constraint on interest. Reputational capital determines who will collaborate on technology, who will sign long-term agreements, and how India's positions are received at the UN and the FATF. A short-term gain that costs standing is a bad trade even in purely strategic terms.
3. Insist on consistency. The test of a principle is whether it is applied when it is expensive. Selective ethics is correctly read as no ethics, and it invites the charge of hypocrisy that India levels at others.
4. Build institutional safeguards so that the outcome does not depend on one official's conscience: inter-ministerial committees, mandatory legal and humanitarian risk assessment, parliamentary oversight, and recorded reasons for every approval and refusal.
5. Insulate the decision from the sales incentive. Those who negotiate a deal should not be those who clear it, and performance appraisal should not reward volume alone.
6. Use conditionality rather than a binary. End-use certificates, no-re-transfer clauses, staged deliveries, verification rights and denial of upgrade support for violations allow a sale to proceed with reduced risk.
7. Be transparent about the framework, discreet about the file. Publish the criteria and aggregate data; protect commercially and operationally sensitive detail. Legitimacy comes from the known rule, not from the disclosed transaction.
8. Retain the capacity to refuse. An export policy with no refusals is not a policy. Ajit's real professional obligation is to ensure that the option of saying no survives contact with the revenue target.

Conclusion. In this case the ethical choice and the strategic choice are the same, which is common and under-recognised. Country A offers a legitimate purpose, a technology partnership and a lower risk of Indian weapons reaching people India is fighting elsewhere. Country B offers cash and an immediate transaction against proliferation risk, diversion risk and a contradiction of India's own declared position on state support to armed groups. Where the two do genuinely diverge, the answer is not to abandon interest but to pursue it within limits fixed in advance, because that is what makes the interest durable.