Editorial 1: Rabies elimination: a death outside the database remains a public-health failure
GS II: Issues relating to development and management of Social Sector/Services relating to Health; GS III: Science and Technology
Context
The 5 October 2026 explanation of India’s rabies statistics confronts a disturbing policy problem: a country seeking to eliminate dog-mediated human rabies by 2030 lacks one comprehensive count of the deaths it is trying to prevent. The Integrated Disease Surveillance Programme-Integrated Health Information Platform reported 54 deaths in 2024. An ICMR-led community study estimated annual mortality at 5,726. These numbers should not be combined or presented as interchangeable measurements. Their difference nonetheless makes it unsafe to treat the small surveillance total as evidence that the public-health problem is nearly solved.
The study surveyed 337,808 people in 78,807 households across sixty districts in fifteen States, then combined community and laboratory information with a probability decision-tree model. The result is an estimate, not a list of certified deaths. Separately, the Medical Certification of Cause of Death system recorded 271 medically certified rabies deaths in 2023. That system covered about 22% of registered deaths nationally, while surveillance has its own reporting network and purpose. A responsible elimination strategy must understand what each source can see before deciding what the apparent trend means.
UPSC Syllabus Mapping
GS II: Issues relating to development and management of Social Sector/Services relating to Health.
GS II: Government policies and interventions for development in various sectors and issues arising out of their design and implementation.
GS III: Science and Technology, developments and their applications and effects in everyday life.
Multi-Dimensional Analysis
1. Count the burden without confusing the instruments
Surveillance, medical certification and community modelling answer different questions. Surveillance tracks cases and deaths reported through a network that can identify unusual patterns and trigger response. Medical certification describes the causes assigned to deaths that receive certification. A community study seeks to estimate population burden, including people who never reach a hospital or whose illness never enters the reporting pathway. None is automatically a complete census merely because its number is nationally published.
The practical consequence is that apparent precision can conceal different coverage. The figures of 21 surveillance deaths in 2022, fifty in 2023 and 54 in 2024 are valuable records, but they are not a basis for asserting that total mortality followed the same path. Similarly, a modelled 5,726 should retain its status as an estimate rather than replace every local record with a fixed national truth. Policymakers need reconciliation through methods, coverage and uncertainty, not a contest over which headline is politically more convenient.
Better reporting can initially increase the recorded count without indicating that disease incidence has worsened. If district staff identify previously missed deaths, the database becomes more complete. A performance system that punishes any rise in reported cases can therefore discourage the very detection needed for elimination. Public dashboards should explain reporting improvements and distinguish them from changes in transmission, while preserving consistent definitions over time.
2. Follow the entire post-exposure care pathway
Rabies illustrates the difference between having a preventive technology and ensuring that an exposed person receives it. WHO describes prompt post-exposure prophylaxis as thorough wound washing, a vaccine course and, where indicated, rabies immunoglobulins or monoclonal antibodies. It advises washing a potentially exposed wound with soap and water for at least fifteen minutes and seeking medical attention. Once the central nervous system is infected and clinical symptoms appear, the disease is virtually invariably fatal. Delay cannot be treated as a minor service inconvenience.
The ICMR-led study’s reported care gaps are therefore operationally significant. About one in five dog-bite victims had not received anti-rabies vaccine. Only 66.2% had received at least three doses, and roughly one in ten eligible victims received passive immunisation. These figures should not be used to prescribe a universal schedule from a newspaper summary. They indicate breaks between exposure, assessment, treatment and completion that trained health services must examine using current clinical guidance.
A clinic may have vaccine on the day of the first visit while the patient still fails to complete care. Travel costs, lost earnings, unclear instructions and difficulty locating the appropriate service can interfere with follow-up. The service unit is consequently the completed care pathway, not the first dose administered. District planning should identify where exposure victims go, what happens if required products are unavailable and how patients are referred without losing critical time.
3. Make One Health a division of responsibility, not a slogan
WHO identifies dogs as responsible for up to 99% of human rabies cases. Human treatment prevents a fatal outcome after exposure, while dog vaccination addresses transmission at its source. A policy that funds only hospital products responds downstream; one that focuses only on animals leaves exposed people without immediate protection. One Health is useful because these tasks are related but cannot be performed by an undefined collective with no accountable owner.
Municipal and animal-health institutions need plans for dog vaccination and bite prevention, while health services need exposure assessment, treatment and reporting. Shared information can show where bites, suspected animal disease and human cases cluster. The purpose is not to turn every complaint into a disease diagnosis, but to connect signals that separate departments would otherwise see in isolation. Local coordination should produce a response timetable and designated responsibilities.
Dog management must also follow evidence rather than public fear. WHO states that culling free-roaming dogs is not effective in controlling rabies and identifies mass dog vaccination as the most cost-effective source-prevention strategy. This does not erase legitimate safety concerns about bites. It requires those concerns to be addressed through vaccination, responsible ownership, prevention education and appropriate local administration instead of treating visible removal as proof that transmission has ended.
4. Notification needs a functioning route from community to district
The 5 October account says notification has historically depended on State requirements and that NCDC has been seeking nationwide notification through the relevant State legislation. A legal requirement can establish an obligation to report, but detection, diagnosis and transmission of the record still have to occur. A death at home cannot enter a database simply because a circular was issued to hospitals. Frontline awareness and clear reporting responsibilities are essential.
IDSP’s suspected, presumptive and laboratory categories illustrate why surveillance need not wait for perfect certainty before a signal is recorded. At the same time, those categories must remain distinct so that an early warning is not silently converted into a confirmed diagnosis. District and State surveillance units should provide training on definitions and feedback on the usefulness of reports. Laboratories and response teams then need the capacity to investigate the relevant signals.
A fragmented reporting arrangement can miss private facilities, referrals and deaths after discharge. Reviewing these paths is more informative than asking only whether every district submitted its monthly form. The institutions must understand where information is lost and repair that point. Families should encounter an accessible public-health process, not several departments each explaining why the death belongs in somebody else’s dataset.
5. Elimination credibility rests on service equity and honest uncertainty
Rabies disproportionately affects people for whom preventive products and travel are difficult to afford. WHO describes it as a neglected tropical disease affecting marginalised populations, while children are frequent victims of exposure. The cost of delayed care therefore includes an equality dimension: the availability of an effective medical intervention means little when a poor family cannot reach or complete it. Elimination must include the populations least likely to appear in routine records.
The current WHO fact sheet uses a model-based global annual-death estimate of 44,203 from the 2021 Global Health Estimates, while older accounts commonly repeat 59,000. This editorial does not blend the two into a fabricated consensus. The example reinforces the larger point that estimates need their date and method. An honest account of uncertainty is a tool for better planning, not an excuse to postpone a known preventive responsibility.
Public authorities should avoid declaring victory from low reported deaths without examining detection and service coverage. Equally, they should not imply that a model proves the exact toll in every district. A credible 2030 programme would demonstrate stronger reporting, effective source prevention and dependable human care. The goal is fewer people dying, even when better surveillance initially makes more of the remaining deaths visible.
Way Forward
1. Reconcile the data streams: Health authorities should publish definitions, coverage and periods for surveillance, medical certification and research estimates. District-level review can investigate mismatches without adding overlapping records or assuming that every difference indicates deliberate concealment.
2. Guarantee a completed exposure-care route: States should map where assessment, vaccine and indicated passive immunisation are available. Referral arrangements and follow-up should be designed around the patient’s next step, so that a missing product does not become an unexplained interruption.
3. Assign One Health owners: District plans should identify the institutions responsible for dog vaccination, bite prevention, human care and case investigation. Joint review should track concrete tasks and local signals rather than record only that departments attended a meeting.
4. Reward detection quality: Surveillance performance should consider completeness, timeliness and follow-through. A rise after reporting improves must be explained and investigated, not automatically treated as failure that encourages officials to keep their numbers artificially low.
5. Strengthen frontline communication: Schools, community workers and clinics should explain immediate wound washing and urgent medical assessment in clear language. Advice must direct people towards trained services and should not suggest that home care or a prior vaccination removes the need for exposure assessment.
6. Test elimination claims independently: Progress reports should examine transmission evidence, treatment access and reporting reliability together. Independent review of hard-to-reach populations can reduce the risk that a quiet database is mistaken for the disappearance of the disease.
Conclusion
India’s rabies challenge is not solved by choosing between 54 and 5,726. It is addressed by understanding why systems see different parts of the burden and then closing the gaps in prevention, care and reporting. A preventable death remains a policy failure even when it occurs beyond the formal health network. The route to credible elimination is a system that can protect the exposed person, interrupt transmission in animals and account honestly for the outcomes it still fails to prevent.
Practice Mains Question
‘Low reported mortality is not sufficient evidence of disease elimination. Examine this statement in the context of India’s rabies surveillance and One Health responsibilities.’ (250 words)
Editorial 2: Legislative privilege: protect the House from obstruction, not the government from criticism
GS II: Parliament and State Legislatures, functioning, powers and privileges; Constitution and fundamental rights
Context
Two reports published on 5 October 2026 present an important procedural qualification to a large constitutional question. An explanation describes the scheduled seven-judge Supreme Court reference on legislative privilege and free speech on 6 October. Another reports that the Tamil Nadu Assembly informed the Court that its disputed 2003 resolution had been withdrawn in 2004, raising a preliminary objection to a merits hearing. The reference may therefore remain formally open without immediate resolution. An exam analysis must not turn a scheduled hearing into an assured final ruling, or treat an oral procedural discussion as a new constitutional doctrine.
The underlying dispute arose after The Hindu criticised the exercise of state power under the Jayalalithaa government. The Assembly sentenced five journalists of the newspaper and Murasoli editor S. Selvam to fifteen days’ imprisonment for alleged contempt and breach of privilege. The Supreme Court intervened against the arrest warrants, but the larger relationship between Articles 194(3), 19(1)(a) and 21 remained unresolved in the reference. The present development invites attention to a principle with implications well beyond that particular newspaper: a legislature needs independence to deliberate, while citizens need freedom to evaluate what it does.
UPSC Syllabus Mapping
GS II: Parliament and State Legislatures, structure, functioning, conduct of business, powers and privileges and issues arising out of these.
GS II: Indian Constitution, significant provisions and fundamental rights.
GS II: Separation of powers between various organs, dispute redressal mechanisms and institutions.
Multi-Dimensional Analysis
1. Functional protection is different from political insulation
Legislative privilege exists because representatives must perform their constitutional work without improper obstruction. Speech and deliberation inside the House need protection from pressures that could undermine representation. But extending that rationale to punish criticism outside the House requires a separate explanation. An unfavourable account can affect the reputation of an officeholder without preventing the legislature from meeting, debating or deciding. Conflating reputational discomfort with institutional obstruction expands a necessary protection into a general power against dissent.
The petitioners and intervenors described in the reporting frame privilege as functional. That is an argument before the Court, not a final new holding. It nevertheless offers a practical test: what legislative task did the impugned act obstruct, and how does the response protect that task? Requiring an answer would not make the House powerless. It would distinguish genuine interference from disagreement, inaccurate reporting that can be corrected, and severe but legitimate political criticism.
The distinction is particularly important where the majority supporting a government also controls a privilege proceeding. Legislatures are not identical to governments, chief ministers or ruling parties. A body that scrutinises the executive should not automatically use its collective powers to protect executive reputation. The democratic purpose of the institution is weakened when the citizen encounters it primarily as an instrument for punishing criticism of those currently governing.
2. Personal liberty requires more than a majority resolution
The 2003 dispute was not only about a newspaper’s editorial freedom. Arrest warrants and imprisonment affected personal liberty. Articles 14 and 21 therefore form part of the petitioners’ account alongside Article 19. A parliamentary vote can be democratically authorised while still requiring examination of whether its use against a person is fair and constitutionally permitted. Constitutional government does not make every act of the elected majority immune from rights scrutiny.
The reported history shows the gravity of this distinction. Police teams were deployed across Tamil Nadu, Karnataka and Delhi to execute the warrants. The coercive state was available to enforce a legislature’s view of contempt against journalists. Where such powers are invoked, notice, a meaningful opportunity to answer and a reasoned account of the alleged obstruction matter. They are protections for institutional legitimacy as well as for the targeted individual.
A person should not have to choose between abandoning scrutiny and risking detention for an undefined offence against legislative dignity. Clear procedure can narrow that fear without denying a legitimate power to respond to actual obstruction. The policy objective is not to substitute courts for every internal legislative decision. It is to ensure that coercion affecting outsiders has a defensible constitutional basis and a reviewable connection to the protected function.
3. Precedent must be read through its dispute and limits
The 1959 Searchlight decision in Pandit M.S.M. Sharma v. Sri Krishna Sinha gave priority to the special privilege provision over general free-speech protection in the dispute before it. The 1964 Keshav Singh reference rejected the reading that all Part III rights must always yield to privilege. The controversy involved an Assembly ordering judges and a lawyer into custody after a citizen obtained bail. This history demonstrates why a simplistic sentence that privilege always wins or never exists would misstate the constitutional development.
Later decisions also supply boundaries rather than one unlimited formula. Raja Ram Pal in 2007 upheld Parliament’s expulsion of eleven MPs arising from a cash-for-query sting while recognising judicial review. Amarinder Singh in 2010 rejected punishment for conduct unrelated to the functioning of the House. The 2024 Sita Soren decision rejected immunity for bribery connected with votes. Each shows that privilege must be assessed in relation to its purpose and the conduct at issue.
The current reference is important because journalists and citizens encounter privilege from outside the institution. Existing decisions about members, votes or expulsion inform the analysis but cannot simply be substituted for the precise question of criticism and reporting. A clear judicial account could identify how functional necessity, free speech, liberty and review fit together. Until then, careful answers must distinguish established holdings from the arguments seeking further clarification.
4. Withdrawal of a resolution does not erase the governance question
The Assembly’s reported statement that the 2003 resolution was withdrawn in 2004 introduces a real procedural issue. Courts may decide that an individual controversy has lost its immediate consequence, particularly when many live cases need their time. The Chief Justice’s reported reluctance to engage in an extended academic exercise should not be misrepresented as an endorsement of the original action. It concerns whether this proceeding is the occasion for deciding the larger question.
At the same time, the petitioners argue that the larger issue remains alive even if the Tamil Nadu incident no longer demands relief. The long interval matters: a dispute referred in 2003 and again to seven judges in 2004 can shape public uncertainty for decades without a final answer. Retaining a case formally open may preserve a route for future consideration, but it cannot by itself give journalists a predictable account of the limits they face today.
Legislatures need not wait for a future judgment to improve their own practice. They can publish procedures, insist on a connection to functioning and use corrections or responses where detention is unnecessary. Such restraint is not surrender to judicial control. It is an institution applying its constitutional purpose to its own powers while acknowledging that lawful criticism remains part of the democratic environment in which it operates.
5. Digital scrutiny makes clear boundaries more urgent
Contemporary reporting circulates through broadcasts, online publications and social media. Criticism can be instant, severe and incomplete, and an incorrect account can spread widely before a correction is issued. These conditions create legitimate institutional concerns, but they also increase the number of citizens who could face expansive privilege claims. A power designed around the functioning of a House should not become a general regulatory tool for every disputed online statement.
Accuracy remains an obligation of responsible journalism. A constitutional defence of scrutiny does not require defending fabricated reports or refusing correction. Yet an error and an obstruction are different matters. Legislatures have public platforms, records and communication channels through which they can clarify proceedings. The choice of remedy should respond to the actual harm rather than assume that imprisonment is the natural expression of institutional authority.
Public trust grows when a House can tolerate criticism while answering it with reasons. Conversely, a severe privilege response can shift attention from the alleged reporting error to the institution’s use of coercion. A functional approach would preserve protection against genuine interference and reduce the chilling effect on those trying to inform citizens. That balance supports, rather than diminishes, the legislature’s standing as a forum of public reason.
Way Forward
1. Publish clear privilege procedures: Houses should make the grounds, stages and possible outcomes understandable to citizens. A person facing a complaint should be able to identify the alleged conduct and the process for answering it without relying on informal access to officials.
2. Require a functional connection: Committees should explain how the conduct interfered with legislative work. A reasoned link would make the power less dependent on the severity of criticism or the political status of the person criticised.
3. Protect liberty through fair process: Where coercive action is considered, notice, hearing and reasoned findings deserve particular care. The response should not proceed merely because a majority regards the publication as offensive or politically inconvenient.
4. Use proportionate correction routes: Legislatures should consider clarification, publication of the record and other suitable responses to reporting disputes. These measures can protect factual accuracy without treating every disagreement as a threat requiring detention.
5. Preserve meaningful review: Institutions should recognise that judicial examination of constitutional limits differs from interference in ordinary debate. Review can protect legitimate privilege by separating its proper exercise from conduct unrelated to legislative functioning.
6. Maintain accurate public communication: Until the reference’s procedural future is settled, reports should distinguish the scheduled bench, the withdrawal submission and any actual order. A constitutional controversy should not acquire a fictitious resolution through an overconfident headline.
Conclusion
Privilege and free speech are not opposing luxuries from which democracy can choose one. A legislature needs freedom to work, while the public needs freedom to examine that work and the government it sustains. The 5 October procedural development may delay authoritative settlement, but it does not remove the duty to use power for its constitutional purpose. Protecting the House from obstruction is defensible; protecting rulers from criticism through the threat of imprisonment is a different proposition that requires the closest scrutiny.
Practice Mains Question
‘Legislative privilege should protect constitutional functioning rather than political reputation. Examine the relationship between privilege, press freedom and personal liberty in India.’ (250 words)
Editorial 3: Voluntary AI safety: an audit promise needs access, independence and consequences
GS II: Governance and accountability; GS III: Science and Technology, developments and their applications
Context
A 5 October 2026 BBC report describes a new United States task force for artificial intelligence after President Donald Trump’s announcement on Sunday. The reported body is intended to coordinate government engagement with consumers, public-interest organisations, infrastructure providers and technology companies. The development follows the voluntary accord announced at the White House on 29 September. The immediate news is therefore a governance response to a prior commitment, not evidence that a legally enforceable safety regime has already replaced voluntary company action.
The Associated Press account carried by PBS describes internal and external review commitments and Trump’s emphasis on self-policing. Council on Foreign Relations analysis published on 1 October questions whether the accord is effective, verifiable, monitorable and enforceable. It describes four proposed layers: internal controls, an internal team, independent external evaluation and independent board oversight. That account supplies a useful analytical framework, but its conclusions are the author’s assessment, not a court ruling or a universally accepted description of future regulatory results. India’s exam-relevant question is what makes an assurance about powerful technology capable of being tested.
UPSC Syllabus Mapping
GS II: Important aspects of governance, transparency and accountability.
GS III: Science and Technology, developments and their applications and effects in everyday life.
GS III: Awareness in the fields of IT and computers; challenges to internal security through communication networks.
Multi-Dimensional Analysis
1. A promise and an obligation produce different incentives
Voluntary commitments can be useful when companies move faster than public institutions can develop detailed requirements. They can identify common concerns and establish practices that were previously left undefined. Yet a commitment’s existence does not establish that a company must bear a consequence for breaking it. The CFR assessment observes that the accord does not have the force of law or an executive order. Its practical strength therefore depends on mechanisms beyond ceremonial agreement.
Commercial incentives matter because resources spent on safety and resources spent on new capabilities compete within the same organisations. A company may sincerely value careful evaluation while also facing pressure to release a product, win customers or justify infrastructure investment. The analytical issue is not whether every executive has bad intentions. It is whether the governance arrangement makes the safer decision viable when its immediate commercial cost is high.
An effective system must define conduct rather than rely only on aspirations. Terms such as independent review or appropriate controls can sound reassuring while leaving each participant free to decide what they require. Clearly specified obligations make it possible to ask what was tested, who saw the results and what happened when a problem was found. Without those questions, the promise can remain intact in public communication even as the underlying practice varies substantially.
2. Independent evaluation requires more than an external name
The accord’s reported external-evaluator layer is potentially valuable because an organisation assessing its own product can have blind spots and conflicts. But independence is not established simply by hiring somebody outside the payroll. The evaluator needs sufficient access to the system, relevant records and the conditions under which it is used. A narrow demonstration selected by the developer may reveal little about failures occurring outside that presentation.
Appointment, remuneration and removal arrangements also influence what the evaluator can say. If continued work depends on satisfying the company whose safety claims are under examination, criticism may carry a professional cost. The CFR author specifically notes missing detail on full access and protection from retaliation. These are design issues for any regime that depends on outside assurance, including standards used in public procurement or critical infrastructure.
Evaluation also needs a defined purpose. Testing whether controls work as intended is different from verifying that an organisation has written a policy. A report should make the scope and limitations visible so that a limited assessment is not marketed as proof of universal safety. Disclosure can protect commercially sensitive information while still stating the nature of the test, its unresolved findings and the conditions in which assurance no longer applies.
3. Board oversight must be connected to authority to change outcomes
An independent committee of directors can place safety concerns at the level where strategic decisions are made. It can ensure that internal staff and outside evaluators have a route beyond a product team that may be focused on delivery. But a board committee is meaningful only if it receives material information and can require corrective action. Merely receiving a periodic presentation does not prove that oversight affected a risky decision.
The reported accord calls for issues to be remediated but does not specify all the conditions of independence or responsibility. This creates an important institutional question: who can stop a release, require additional testing or restrict use after a serious finding? The answer should be known before an incident. Otherwise a concern can circulate between teams while each believes that another group holds the final responsibility.
Oversight also has to preserve dissent. Technical staff may identify weaknesses that are difficult to communicate in executive language, particularly when the system performs well in most routine cases. Channels for escalation should let those concerns reach decision-makers without being reduced to an average performance score. A rare high-consequence failure can remain important even when a product’s overall success rate looks attractive.
4. Safety is contextual, especially in public and infrastructure use
The same model can create different risks depending on where it is deployed and what authority it is given. A tool used to suggest draft text differs from one able to affect operational infrastructure or make consequential decisions. A broad safety label cannot settle those differences. Public institutions should assess the actual use case, information exposure and ability to recover from a mistaken output, rather than rely only on a provider’s general claim.
The BBC report’s description of engagement with consumers and critical infrastructure providers makes this distinction relevant. Those groups face different consequences and need different evidence. Consumer communication may emphasise limitations and recourse, while infrastructure use demands stronger testing of failure conditions and continuity. A single voluntary statement can establish a discussion framework but cannot substitute for these specific responsibilities.
India can draw lessons without mechanically importing the reported American arrangement. Public-service adoption should identify the accountable department and retain a route for human review where a decision affects a person. Technology can support administration, but a person denied a service needs an explanation and remedy from the institution responsible. Procurement should consequently examine operational limits and correction arrangements alongside price and performance.
5. Coordination should reduce risk without disguising market power
The accord’s promise of regular meetings to develop standards can support information sharing about safety. Companies may learn about failures that others have already encountered and avoid repeating them. At the same time, collaboration among major firms raises questions about who defines acceptable practice and whether smaller participants or public interests are represented. Industry coordination should not quietly become the sole source of rules for the industry itself.
Transparent participation and a defined scope can help separate safety cooperation from commercial coordination. Public authorities need enough technical capacity to interrogate the proposals rather than depend entirely on the firms’ descriptions. External researchers, affected groups and smaller developers can bring evidence about risks that a meeting of the largest producers may overlook. Broader engagement is useful when it has a route into decisions, not merely an invitation to attend.
Naming a new task force does not resolve these issues by itself. Its remit, information access and outputs determine whether it changes governance or only coordinates communication. Current reporting establishes an announcement, while the effectiveness remains a future question. A careful public response should welcome useful institutional work without attributing enforceable powers or successful oversight that the sources do not demonstrate.
Way Forward
1. Specify the commitment: Safety arrangements should identify required tests, responsible institutions and the conditions for release or use. Clear conduct can be assessed; general assurances cannot reliably show whether the promise was kept.
2. Design genuine evaluator independence: External reviewers need appropriate access, protection for adverse findings and a scope that is not selected solely to display success. Reports should disclose limitations so that narrow testing is not mistaken for complete assurance.
3. Give oversight a decision route: Board and institutional committees should know how a concern can delay, restrict or stop a deployment. Escalation needs an accountable destination before a technical warning encounters commercial pressure.
4. Match safeguards to use: Public procurement should examine the consequences of error, available recovery and authority given to the system. High-stakes use needs evidence and recourse suited to that setting rather than a generic assurance inherited from another application.
5. Preserve public participation: Standards discussions should include relevant researchers, users and affected groups, with a clear record of how their concerns were considered. This can improve legitimacy while avoiding the assumption that large-company agreement is equivalent to the public interest.
6. Review incidents and outcomes: Authorities should evaluate whether controls actually detected and corrected failures. A governance framework deserves revision when evidence shows that written procedures failed in operation, even if every participant can show that it signed the original accord.
Conclusion
Voluntary AI commitments can begin a useful governance conversation, but safety cannot be inferred from signatures or the appointment of a coordinating body. It needs clear responsibilities, evidence that independent reviewers can test, and consequences when a serious concern is ignored. The current American developments offer India an institutional lesson rather than a finished template: innovation and accountability become compatible when the system identifies who must act, what they must demonstrate and how those affected can obtain a correction.
Practice Mains Question
‘Voluntary technological-safety commitments are credible only when they are verifiable and connected to accountability. Discuss with reference to AI evaluation, board oversight and public-sector adoption.’ (250 words)
