Editorial 1: Rebuilding examination trust: computerisation is a change in risk, not the end of risk
GS II: Education, governance and accountability
Context
The Department-related Parliamentary Standing Committee on Education, Women, Children, Youth and Sports presented its 382nd report on reforms and strengthening of the National Testing Agency on 9 October 2026. Chaired by Mukul Wasnik, the committee had adopted the report on 25 August after hearing ministries, NTA officials and the high-level expert committee chaired by K. Radhakrishnan. Its recommendations should be read as an agenda for institutional change, not as rules that have already been imposed on every examination.
The report begins from the scale of NTA’s responsibility: about 1.2 crore students sit its examinations annually. It administers gateways to higher education such as JEE-Mains, NEET-UG, UGC-NET and CUET, making examination reliability a matter of public opportunity rather than an internal logistics issue. When that gateway fails, the consequences reach students, families, institutions and public administration. Restoring trust requires an account of who is responsible at each stage and whether the process tests the learning it claims to assess.
Computer-based testing is central to the proposed response, but the report itself refuses the simple equation of digital with secure. It seeks government-owned infrastructure, question-bank reform, transparent normalisation, an equity audit and in-house cybersecurity. It also notes the influence of coaching on examination design. These recommendations together describe a wider institutional problem: equipment and surveillance can support a test, but they cannot independently establish fair access, sound assessment or accountable decision-making.
UPSC Syllabus Mapping
GS II: Issues relating to development and management of Social Sector/Services relating to Health, Education, Human Resources.
GS II: Important aspects of governance, transparency and accountability, e-governance.
GS II: Parliament and State legislatures – structure, functioning, conduct of business, powers and privileges.
Multi-Dimensional Analysis
1. Question security requires a different production process
The committee identifies investment in a large, diverse and continuously updated question bank as a long-term defence against paper leaks. Its proposed mechanism separates the creation of individual questions from the final selection of a paper. Questions would be drawn using digital tools close to the examination and vetted by experts who remain sequestered until it ends. That design aims to reduce the number of people who can know the complete final paper far in advance.
This is more specific than calling for stronger security. It changes when sensitive information exists, who can access it and how the paper is assembled. A secure question bank still requires control over permissions, version history and access records. Otherwise risk may move from printed-paper transport into the digital repository. The value of the proposal lies in reducing exposure while recognising that the new process also creates a system that must be governed.
Expert integrity is another part of the same chain. The report calls for checking conflicts of interest among contributors, including links to coaching centres and relatives taking the examination. Institutional reputation alone does not replace that check. A trusted academic can still have a relevant conflict, while a clear screening process protects both the institution and the legitimacy of the test. Security therefore includes professional governance as well as technical controls.
2. Digital transition must not shift the burden to disadvantaged candidates
The report recommends moving NEET-UG to computer-based testing but links that transition with prior investment in public infrastructure. It supports a phased plan for 1,000 secure government CBT centres, at least one test centre per district and possible mid-sized school or college IT labs with 100-200 nodes. These are proposals for capacity, not a count of centres already built. The sequencing matters because a new test format can otherwise impose travel and accommodation costs on students who already face access barriers.
Mock examinations are also recommended so candidates can become familiar with the interface. Familiarity is not a minor convenience in a high-stakes assessment. If students differ in their ability to operate the test interface, the result may partly measure exposure to computers rather than the academic construct intended. Practice opportunities should therefore be accessible enough to reduce that unintended source of difference, not merely available to those already comfortable online.
Government ownership is presented as a way to address weaknesses associated with subcontracted private centres. Ownership can clarify responsibility, but public status alone does not establish secure operation. Staff, power, connectivity and maintenance must support the facility. The committee’s infrastructure proposal should be assessed through readiness and access evidence rather than a ceremonial count of sites inaugurated. A transition becomes fair when the candidate’s practical journey works, not when the procurement file closes.
3. Fair scores require transparent assessment, not only uniform software
Multi-session examinations create a normalisation problem when candidates sit different papers. The committee calls for a transparent, independently audited methodology to be developed and explained before any CBT transition. That requirement recognises that a score becomes consequential through its relation to other scores. A technically functioning examination can still produce contested ranks if candidates do not understand how differences in paper difficulty are handled.
The report also criticises an assessment design that rewards coaching-specific rapid recall and pattern recognition over conceptual understanding based on the official curriculum. This shifts attention from test administration to the purpose of the test. An examination can be leak-free and still favour candidates who can buy specialised preparation for material or techniques beyond school learning. Reform needs measurable standards for what the paper should assess, not an assumption that tougher questions are automatically better questions.
The proposed research wing would give NTA capacity to examine whether its tests fulfil their intended purposes over time. That is an institutional answer to a recurring weakness: administering an examination is different from validating it. Research can investigate relationships between scores, access and subsequent educational outcomes. Such work should inform design without being used to promise that one test can capture every form of ability or eliminate all inequalities in education.
4. Accountability needs equity evidence and a visible vendor record
The committee observes that participation data and success data answer different questions. A State can have many registered candidates while relatively few succeed, potentially pointing to preparation or access problems. It proposes an annual Equity Audit by State, gender, school-board type, medium of instruction and social category. This would turn a general claim about fairness into evidence about where differences occur and which groups need closer examination.
An audit does not automatically identify the cause of every gap. Differences may arise before the examination as well as within its design and administration. The value of disaggregation is to locate questions that require explanation, not to infer discrimination from every unequal percentage without investigation. Published methods and responsible interpretation are necessary so useful evidence does not become a simplistic ranking of communities or school systems.
Vendor accountability is another proposed safeguard. The committee recommends a National Registry of Blacklisted Examination Vendors, covering associated individuals and entities as well as firms. The problem is that exclusion by one authority may not prevent the same network from contracting elsewhere. A shared record can reduce that gap, but entries must be accurate and maintained. A registry with outdated or poorly attributed information could either miss relevant risks or unfairly exclude a supplier without a clear basis.
5. Institutional capacity determines whether reform survives the next crisis
The report discourages additional-charge appointments to key NTA positions and notes that even a full sanctioned strength of 39 officers leaves reliance on contractors. It seeks greater internal logistical capacity and a dedicated cybersecurity team able to address remote access, screen recording and virtual-machine risks. These are organisational requirements, not simply a request to purchase another layer of monitoring equipment.
The committee also records an accumulated revenue surplus of Rs 590 crore between 2018-19 and 2024-25 and argues for investment in examination infrastructure. A surplus should be examined in relation to the agency’s public responsibility, while any spending proposal still requires sound design and accountability. The question is whether available resources strengthen the reliability and accessibility of the service students depend on, rather than whether a large balance by itself demonstrates good administration.
Finally, the report recognises the mental pressure associated with high-stakes examinations and recommends accessible support. A calm testing environment and clear procedures matter alongside integrity controls. Surveillance that treats every candidate as a potential offender can conflict with a supportive experience if poorly designed. Institutional reform should protect honest students while investigating malpractice, maintaining the distinction between necessary security and avoidable disruption of the very people the examination serves.
Way Forward
1. Sequence the digital transition: The Ministry and NTA should establish public-centre readiness and accessible mock-test arrangements before changing NEET-UG’s format. Announcing a transition is not evidence that its access conditions have been met.
2. Govern the question-bank chain: Expert screening, controlled permissions and final-paper vetting should be designed together. Moving content into a digital bank must reduce exposure rather than relocate an undocumented risk.
3. Publish assessment methods: Normalisation should be explained and independently audited before use. Curriculum alignment and conceptual assessment need measurable standards that can be reviewed after each examination.
4. Turn equity data into inquiry: The proposed audit should publish disaggregated participation and success evidence with methods. Findings should guide investigation and educational support, not unsupported causal claims.
5. Build durable internal capacity: Full-time leadership, examination research and cybersecurity expertise should have clear responsibilities. Vendor records and contracts need maintenance and review beyond the crisis that prompted them.
6. Evaluate the complete student journey: Centre access, procedures, results and support should be assessed together. The purpose is an examination that is secure, understandable and fair in practice, not only successfully delivered by software.
Conclusion
The 382nd report’s strongest contribution is its refusal to reduce trust to one device or procedure. Computerisation can change the security chain, but it also demands infrastructure, audited assessment and institutional capacity. NTA reform should therefore be judged through the experience and opportunities of students, the quality of the test and the credibility of accountability. Recommendations become public value only when their sequence, resources and outcomes are made visible.
Practice Mains Question
‘Computerisation of public examinations changes the risk structure without automatically ensuring fairness.’ Examine the roles of infrastructure, assessment design, equity audits and institutional accountability. (250 words)
Editorial 2: Stubble burning: change the farm decision, and measure the result honestly
GS III: Agriculture, environment and policy incentives
Context
A 9 October 2026 analysis by economist C.S.C. Sekhar in The Hindu examines why paddy-stubble burning returns despite years of interventions. It links the practice with groundwater law, the paddy-wheat cycle, procurement incentives, machinery costs and limitations in satellite measurement. The argument is important because it treats burning as a farm decision made under constraints rather than a problem that can be solved simply by adding another prohibition.
The article estimates that Punjab and Haryana together produce about 29 million tonnes of paddy straw. It describes a narrow 10-15-day interval between delayed paddy harvest and wheat sowing, during which residue must be managed. These numbers are attributed to the analysis, not a fresh nationwide survey independently conducted here. They nevertheless help identify the scale and timing of the practical task a farmer faces. A solution that arrives after the sowing window has closed is not an effective alternative, even if technically sound.
The same analysis questions a reported 90% fall in fire incidents between 2022 and 2025 by pointing to differences between active-fire detection and burned-area evidence. Measurement is therefore part of the policy problem. If the indicator fails to observe activity shifted outside satellite passes, apparent improvement can diverge from conditions on the ground. Reform must change incentives and operations while ensuring that its evidence can distinguish actual reduction from changes in observation.
UPSC Syllabus Mapping
GS III: Major crops, cropping patterns in various parts of the country, different types of irrigation and irrigation systems.
GS III: Issues related to direct and indirect farm subsidies and minimum support prices.
GS III: Conservation, environmental pollution and degradation, environmental impact assessment.
Multi-Dimensional Analysis
1. Water conservation and crop timing create a policy interaction
The article links later paddy transplantation with the Punjab Preservation of Sub-Soil Water Act, 2009. The law sought to reduce groundwater depletion by aligning cultivation more closely with monsoon conditions. That objective addresses a real resource problem, but its interaction with the later wheat crop can compress the interval available for residue management. Examining that interaction does not mean groundwater protection should be abandoned; it means policy must recognise effects across the crop calendar.
The account identifies long-duration PUSA-44, taking about 150-160 days, as an additional pressure on that interval. Combine harvesting leaves residue while farmers need to prepare the field quickly. The resulting decision is shaped by the technology of harvest, the variety planted and the time remaining. A farmer may be asked to avoid burning without receiving an alternative that can operate within those same constraints. Enforcement alone cannot supply a machine or create another week in the calendar.
Short-duration varieties and changes in cultivation practice can reduce pressure, but their suitability must be assessed locally. Yield, labour, water and market expectations all influence adoption. Presenting a different variety as a universal cure would repeat the mistake of ignoring how several policies and farm choices interact. A credible approach tests alternatives within the actual production sequence, not as isolated interventions whose benefits are assumed to add together automatically.
2. Procurement certainty makes diversification a risk decision
The analysis compares average paddy procurement between 2012 and 2024 at about 37% of production nationally, 87% in Punjab and 74% in Haryana. For wheat it gives about 30% nationally, 70% in Punjab and 63% in Haryana. These are historical averages in the author’s account, not new 2026 season rates. They show why farmers can perceive rice and wheat as more dependable sources of income than crops with less assured purchase.
Diversification therefore asks the grower to change both agronomy and market exposure. A recommendation to plant another crop may be environmentally appealing while leaving unanswered who will buy it, at what quality standard and how payment will arrive. Assured procurement for familiar crops can outweigh a general promise that alternatives are profitable. Policy needs to address the relative risk faced by the farmer rather than assume that information alone will overcome that difference.
The article also identifies subsidised power as part of the incentive to grow water-intensive crops. Altering one element without considering the others can lead to disappointing adoption. A diversification strategy should examine demand, water use, input cost and income stability together. The aim is to build a viable pathway away from pressure on soil, water and air, not to transfer environmental responsibility to a farmer while leaving the economic structure unchanged.
3. Residue technology works only when the service arrives in time
In-situ approaches keep and manage residue in the field, while ex-situ approaches remove it for uses such as bioethanol or compressed biogas. The distinction is operationally useful because the two routes require different equipment and coordination. A machine suited to sowing through residue does not create a transport and storage system for material leaving the farm. Each route needs its own workable service chain.
Sekhar’s account describes expensive machinery, hiring queues and tractors of around 60 horsepower as constraints for some in-situ equipment. It quotes hiring costs around Rs 10,000 and ex-situ costs of Rs 1,500-2,500 per acre. These figures should be understood as the author’s reported costs, not universal tariffs. Their policy relevance is that an option can exist in a catalogue while remaining unaffordable or unavailable when many farmers need it simultaneously.
A subsidy measured by machines purchased may consequently overstate usable capacity. The questions are where the machines are, who can hire them, whether operators are available and whether a field can be served within the sowing window. Ex-situ users likewise need dependable collection and payment arrangements. A plant announced as a destination for straw does not guarantee that material is collected from dispersed farms at a cost the farmer can accept.
4. Incentives and enforcement should be judged through the feasible alternative
The article contrasts approaches that rely heavily on machinery with a Haryana package combining incentives for residue management, direct seeding, alternative crops and panchayat performance. It cites Rs 1,000 per acre for CRM, Rs 4,000 for direct seeding and Rs 7,000 for diversification. These are reported policy amounts in the analysis, not evidence that every eligible grower has received payment or that each measure is equally effective in every setting.
The point of comparison is institutional design. An incentive can help make a lawful alternative more attractive, but payment timing, eligibility and the cost being offset determine its practical value. A nominal amount may leave an important cost uncovered or arrive too late to influence a seasonal decision. Evaluation needs the farmer’s realised experience rather than only the scheme’s stated benefit. This is also why a policy package should not be declared successful solely because it contains several interventions.
Enforcement remains relevant to harmful conduct, yet its credibility depends on the availability of feasible alternatives and accurate evidence. The analysis warns of incentives to display reduction without achieving it. A system that rewards officials only for lower recorded counts can create pressure to manage the indicator rather than the problem. Accountability should examine service provision, implementation and environmental results together, keeping penalties distinct from claims about the overall success of a programme.
5. Satellite indicators need interpretation and independent checks
The inspected account explains that MODIS and VIIRS-based active-fire observations see India at particular daytime passes and detect fires active when observed. They do not directly count every fire throughout twenty-four hours. That does not make satellite monitoring useless; it defines what the indicator can measure. If burning shifts in time, the relationship between the observed count and actual activity can change.
The article cites an alternative burned-area assessment showing a more gradual reduction of around 30%, compared with the reported 90% reduction in incidents. These figures concern different measurements and should not be averaged into a compromise statistic. Their disagreement is a reason to examine methods, timing and denominators. A responsible analyst preserves the distinction and seeks corroboration rather than selecting whichever number best supports a preferred policy narrative.
Using polar and geostationary evidence, field checks and burned-area information can improve the picture. No single indicator necessarily answers every question about frequency, area and pollution contribution. Measurement design should specify the policy question first. This also supports fair enforcement: a reliable evidentiary chain is necessary when observations may affect farmers or officials, while regional trends require a method capable of seeing changes beyond one satellite’s sampling window.
Way Forward
1. Coordinate the crop calendar: Agriculture and water authorities should assess variety, transplantation and harvest timing together. Water protection needs a residue-management pathway that fits the next crop’s practical requirements.
2. Make diversification economically credible: Procurement and market support should address demand and payment risk for alternatives. Advice alone cannot offset a dependable rice-wheat income pathway.
3. Measure usable machinery service: Programme monitoring should track location, operators, hiring and timely field coverage. Equipment ownership is an input, not proof that farmers received an alternative when needed.
4. Build complete ex-situ chains: Collection, transport, storage and reliable purchase arrangements should accompany straw-using facilities. Announced processing capacity cannot substitute for a workable farm-level route.
5. Evaluate realised incentives: Authorities should examine eligibility, payment timing and the costs actually offset. Support should be assessed through changed decisions and outcomes rather than the number of scheme labels.
6. Use complementary evidence: Active-fire counts should be read alongside burned area, timing and field validation. Public reports must distinguish metrics and explain why different estimates may diverge.
Conclusion
Stubble burning sits at the intersection of the farm calendar, income security, water policy and service availability. Its persistence cannot be explained by one cause or solved through one device. A workable response must change the decision available to the farmer and then measure the resulting change honestly. That means maintaining groundwater protection, building credible alternatives and resisting the temptation to treat a falling indicator as sufficient evidence that the underlying problem has disappeared.
Practice Mains Question
‘Crop-residue burning is an incentive and coordination problem as well as an environmental offence.’ Analyse the roles of crop calendars, procurement, machinery services and measurement. (250 words)
Editorial 3: AI for development: adaptation and public capability matter more than the largest model
GS III: Technology, development and economy; GS II: Public services
Context
The World Bank’s World Development Report 2026, The Promise of Artificial Intelligence, was launched in India on 8 October 2026. Its central framework is adopt, adapt and advance: use existing tools, fit them to local needs and build more advanced capability as resources and institutions permit. The report challenges the idea that every developing economy must begin by building the largest frontier model. It also rejects the opposite assumption that importing a tool is sufficient to obtain its development benefits.
The Bank describes three features shaping the opportunity: capabilities, concentration and complements. AI can expand access to scarce expertise, but advanced production is concentrated in a small number of firms and economies. Its usefulness depends on electricity, connectivity, education, data and institutions. These are not secondary details to fix after deployment. They determine whether a tool can perform a useful task for the people and organisations expected to use it.
The report notes that middle-income countries accounted for half of ChatGPT’s global traffic within six months of launch. That is evidence of rapid diffusion, not proof of equivalent productivity gains or public-service improvements. The distinction between use and benefit is central to responsible policy. A development strategy must explain which constraint AI addresses, who can access the result and how performance will be evaluated, rather than treat adoption statistics as an outcome in themselves.
UPSC Syllabus Mapping
GS III: Science and Technology – developments and their applications and effects in everyday life.
GS III: Indian Economy and issues relating to planning, mobilisation of resources, growth, development and employment.
GS II: Important aspects of governance, transparency and accountability, e-governance applications, models, successes, limitations, and potential.
Multi-Dimensional Analysis
1. Scarce expertise is the practical starting point
The report highlights health screening, agricultural advice and educational support as areas where AI can extend expertise. This is a more useful starting point than asking only how many models or data centres a country owns. A technology has development value when it helps address a constraint that people actually face. The relevant evidence is whether the service works in its setting, not whether the application can produce an impressive demonstration.
The Bank emphasises low-cost “small AI” that can operate through messages, voice calls and basic phones. This matters because a service designed around expensive devices or constant broadband can exclude the very users who lack expert support. Delivery design is part of technological effectiveness. A capable tool that cannot reach its intended audience is incomplete, while a narrower tool on an accessible channel may solve a meaningful local problem.
The expansion of expertise nevertheless requires a boundary around responsibility. Support for a teacher, clinician or official is different from replacing the professional’s judgment in every situation. Public deployment should specify what the system does, when human review is required and how errors are handled. The opportunity is to improve capacity, not to declare a workforce shortage solved merely because an automated interface has been installed.
2. Adaptation is a substantive process, not a translation label
The report argues that adopting a tool developed in a high-income setting can produce recommendations that do not fit local norms and constraints. Language is one part of that mismatch, but institutions, service availability and user needs matter too. A recommendation can be linguistically clear while referring to a service the recipient cannot obtain. Local adaptation must therefore examine the complete route from advice to action.
Data also require preparation and governance. Public administrative records may contain useful information without being ready for safe and accurate AI use. Definitions, missing entries and permissions affect what a system can infer. Making data available indiscriminately is not the same as making it suitable. The report’s emphasis on AI-ready administrative data should be read alongside the need for clear responsibilities over access and quality.
Adaptation can create opportunities for domestic firms and researchers without requiring them to reproduce every layer of frontier production. Local expertise can identify errors, build interfaces and evaluate use cases. That work deserves attention because it links technology with the conditions that determine benefit. A policy that measures innovation only by model size could neglect the smaller but necessary work through which citizens actually receive a useful service.
3. Concentration creates dependency, while full autonomy has a cost
The Bank describes the AI value chain as concentrated in a few economies and companies across chips, infrastructure and advanced models. This creates dependency risks for countries using those capabilities. Supplier choices can affect access, prices and the suitability of available tools. Recognising that concentration is necessary, but it does not mean every economy can economically replicate the entire chain at once.
The report notes that advancing frontier AI requires specialised chips, large data centres, extensive training data and highly skilled researchers. These requirements make it the hardest and most expensive pathway. Pursuing it without examining opportunity cost can divert resources from public foundations or applications with more immediate benefits. Strategic ambition should therefore be sequenced rather than confused with a universal mandate to build the largest system.
Interoperability and the ability to work with multiple providers offer another way to manage dependency. The report advocates national systems that do not lock themselves into one supplier. Procurement and architecture need to preserve that option in practice, including the ability to move or continue a service when arrangements change. Sovereignty is thus partly an institutional capacity to choose and evaluate, not solely ownership of every piece of infrastructure.
4. Government buying power needs testing and staff capability
As users of AI, governments can shape markets through procurement in education, health and administration. The Bank stresses testing, evaluation and scaling successful uses rather than moving directly from a pilot to universal rollout. A public demonstration often operates with extra support and a restricted task. Performance at that stage should not be assumed to continue across different districts, languages or workload levels without further evidence.
The report identifies limited evidence for selecting and evaluating citizen-facing applications as a constraint. Buying a system therefore requires public officials able to define the service objective, assess claims and monitor the result. A vendor contract cannot supply all of that judgment. Government capacity is a complement to the technology because someone must determine what acceptable performance means and whether the service remains within that boundary.
Procurement should also preserve remedies for users. An error that changes access to a service is more consequential than a failed demonstration. People need understandable explanations and a route to seek correction, while officials need records adequate to investigate. This is an analytical requirement flowing from high-stakes use, not a claim that the report has already created one binding global redress mechanism. Local legal duties and administrative responsibilities remain relevant.
5. Employment and power effects require more than a growth forecast
The report distinguishes the immediate effect in economies with many manual workers from disruption in knowledge-intensive services. It expects many early uses to complement workers, while acknowledging job disruption in some sectors. This is a conditional assessment, not a guarantee that no worker will be displaced. Policy must examine tasks, skills and local industry rather than apply one global employment claim to every occupation.
The Bank also notes that AI can widen productivity gaps when complements are absent. Firms with reliable infrastructure and skilled staff may benefit faster than others. A technology that expands capability can therefore increase inequality if access and institutions develop unevenly. Education, business conditions and affordable infrastructure remain part of the economic response, even where model capabilities advance rapidly.
Power effects extend beyond employment. The report discusses concentrated corporate influence, surveillance and propaganda as risks alongside public-service gains. Its recommended regulatory starting points include voluntary standards and existing laws, followed by assessment of gaps and enforcement capacity. This is the Bank’s policy recommendation, not proof that voluntary standards always suffice. Governments must judge whether their legal and institutional arrangements protect people in the uses actually deployed.
Way Forward
1. Begin with a defined constraint: Public agencies should identify the service or expertise gap before choosing an AI product. Evaluation should ask whether the tool improves that task for the intended users.
2. Invest in local adaptation: Language, data and institutional fit should be tested together. Advice must connect with services and conditions that exist in the user’s setting.
3. Build the complements: Electricity, connectivity, education and administrative capacity deserve sustained investment. A model cannot compensate automatically for the foundations its use requires.
4. Preserve supplier choice: Architecture and procurement should support interoperability and assess dependence. Developing local capability should include the ability to evaluate and move between providers.
5. Scale only after evidence: Pilots should be reviewed across relevant conditions before expansion. Staff need the skills and records required to supervise performance and handle errors.
6. Assess distribution and rights: Employment effects, privacy and access should be examined alongside productivity. Regulation should respond to actual harms and institutional gaps rather than rely only on adoption enthusiasm.
Conclusion
The World Development Report offers a practical alternative to the contest over who owns the largest model. Development gains depend on fitting AI to local constraints and building the institutions that make it useful and accountable. Adoption opens an opportunity, adaptation determines much of its value, and advancement requires a realistic assessment of resources. The public objective should be stronger capabilities and better services, measured through results rather than technological scale alone.
Practice Mains Question
‘For developing economies, adaptation and institutional complements are more important than frontier-model ownership in realising AI’s benefits.’ Critically examine. (250 words)
