Chapter 7: The Screen and the Chalk

---

In February 2023, at an Odisha Adarsha Vidyalaya in Koraput district, a Class 9 student named Subhashree stood in front of a sixty-five-inch interactive smartboard and presented her science project on photosynthesis. She had prepared her slides using the school’s computer lab. She referenced a video from the Samagra Resource Portal. Her teacher — one of the OAV’s dedicated, full-time, CBSE-trained faculty — guided her through the question-and-answer session that followed. The classroom had electricity, ceiling fans, a functional projector, broadband internet, a whiteboard, and twenty-eight students sitting on chairs at desks. After class, Subhashree walked to the school library — small but stocked — and borrowed a book on Indian scientists. She was fourteen. She was the first person in her family to use a computer. She was two years away from sitting for the CBSE board examination. Her school was one of 315-plus Odisha Adarsha Vidyalayas built since 2015, residential institutions modeled on the Navodaya Vidyalaya concept, with entrance-exam-selected students, dedicated teachers, functional hostels, and a CBSE curriculum delivered in English medium. The OAV network was designed to do exactly one thing: prove that government schools in Odisha can produce outcomes comparable to the best private schools, if the institution is built right.

Ten kilometers away, at a government primary school in a tribal hamlet, a different scene. The school had one teacher for forty-seven students across Classes 1 through 5. The teacher had a smartphone — a Redmi purchased with his own money — on which the DIKSHA app was installed. The app was installed because a circular from the block education office had required it. He had attended a one-day training session on DIKSHA usage nine months earlier, conducted in Odia, for a platform whose interface was partly in English. He had opened the app three times since training. On two of those occasions, the content did not load because the school had no internet connection and the downloaded modules had been cleared from his phone to free storage space for WhatsApp. On the third occasion, he played a Class 3 mathematics video for seven minutes, during which twelve students watched and the remaining thirty-five did other things, because there was one phone and the screen was five inches. The school had electricity on paper. In practice, the power supply was erratic — available roughly four hours a day during winter, less during monsoon when lines frequently went down. There was no projector, no computer, no library, no internet connection. The students sat on a concrete floor. The drinking water came from a borewell. The toilet for girls existed but had no water connection.

The distance between these two schools is ten kilometers. The distance between the educational experiences they provide is a generation.

This chapter is about the promise that technology can bridge that gap — and the evidence, accumulating across India and worldwide, about when that promise holds and when it does not. The cross-domain question is about leapfrogging: whether digital technology can allow Odisha’s education system to skip the slow, expensive process of building functional institutions, the way mobile phones allowed developing countries to skip landline telephony and mobile payments allowed them to skip branch banking. The answer matters because it determines whether the state invests primarily in screens or primarily in the institutions behind them.

[Note: Subhashree’s school is based on documented OAV specifications and publicly available data on OAV infrastructure and curriculum. The primary school description is a composite drawn from UDISE+ infrastructure data (2,182 schools without electricity, 38% without tap water, 75% without computers), CAG audit findings (43% of test-checked schools with students on floors), and documented DIKSHA adoption patterns. The specific details are representative of conditions documented across tribal district primary schools. Confidence in the pattern: very high, ~90%. Confidence in individual attributed details: moderate, ~70%.]

---

The Leapfrog Question

In 2007, M-Pesa launched in Kenya. Within five years, it had more users than Kenya had bank accounts. By 2024, it processed transactions worth over $314 billion annually across seven countries. The logic was elegant: Kenya had approximately 2,000 bank branches serving forty million people. Building enough branches to provide universal banking would have taken decades and billions of dollars. But Kenya had rapidly growing mobile phone penetration. Safaricom — the mobile operator — realized that the phone could replace the branch. You did not need to build a physical institution if you could deliver the function digitally. Voice communication had already leapfrogged: Kenya went from minimal landline penetration straight to mobile phones, skipping the entire century-long infrastructure build that developed countries undertook. M-Pesa did the same for money transfer. The function — moving money from person A to person B — was fully replaceable by the digital tool. The institution — the bank branch with its tellers, safes, and ledgers — was not needed because the function could be performed without it.

India’s own leapfrog story follows the same logic. UPI replaced the bank transfer queue. Aadhaar replaced the ration card verification chain. JIO replaced the broadband infrastructure that would have taken decades to wire across rural India. In each case, the digital tool replaced the function entirely. You do not need a bank branch to transfer money if you have UPI. You do not need a landline cable to access the internet if you have 4G. The institution was skipped because the function did not require it.

Now apply this logic to education. Can DIKSHA replace the teacher? Can a smart classroom replace a functional school? Can AI tutoring replace years of structured cognitive development guided by a trained adult?

The mobile phone leapfrog worked because voice communication is a function. One person speaks, another hears. The channel — copper wire, radio wave, fiber optic cable — is irrelevant to the function. The phone replaced the channel, not the communication. M-Pesa worked because money transfer is a function. Debit one account, credit another. The mechanism — paper ledger, bank clerk, mobile code — is irrelevant to the function. The app replaced the mechanism, not the transfer.

But education is not a function. It is three things simultaneously: a relationship (student and teacher, built over time through trust, challenge, feedback, and example), an environment (peers who compete and collaborate, a library that invites exploration, a laboratory that makes abstraction tangible, a playground that builds the body alongside the mind), and a process (years of structured cognitive development where each stage depends on the previous one, where gaps in early foundations propagate through every subsequent year). Technology can enhance all three. A smartboard makes the teacher’s explanation more vivid. An online library extends the environment beyond the physical shelves. An adaptive learning platform adjusts the process to the individual student’s level. But technology cannot replace any of the three. A smartboard without a competent teacher is a television. An online library without reading habits is a server full of files nobody opens. An adaptive platform without a trained facilitator is software running on hardware that has no user.

This is the leapfrog fallacy applied to education: the assumption that because digital technology replaced the institution in other domains, it can replace the institution in education. The evidence from the past decade — including from Odisha specifically — suggests that technology in education follows a different rule. It is a multiplier, not a substitute. It multiplies the effectiveness of a functional institution. Applied to a broken institution, it multiplies zero.

---

The COVID Revelation

The pandemic did not create India’s digital education divide. It revealed it, the way an earthquake does not create structural weakness in buildings but reveals which ones were built on sand.

On March 24, 2020, Prime Minister Modi announced a nationwide lockdown with four hours’ notice. Schools closed. They would remain closed, with intermittent and partial reopenings, for nearly two years in most Indian states. Odisha’s schools were closed from March 2020 through February 2022, one of the longest closures globally. The World Bank estimated that India’s school closures were among the most prolonged anywhere — longer than most of sub-Saharan Africa, longer than most of Latin America, longer than virtually all of East and Southeast Asia.

The government’s response was digital. PM eVIDYA was launched in May 2020, integrating DIKSHA (for school education) and SWAYAM (for higher education) with dedicated TV channels (SWAYAM PRABHA) and radio broadcasts. The theory was that digital infrastructure could substitute for closed school buildings. Content would flow from centralized platforms to individual homes via phones, tablets, and televisions. Learning would continue. The institution was closed, but the function would be delivered digitally.

What actually happened in Odisha — and in Bihar, Jharkhand, Chhattisgarh, and Uttar Pradesh — was this: over eighty percent of government school students received no instructional resources to study from home during lockdown. Not inadequate resources. No resources. The number is worth pausing on. Eight out of ten children in government schools — which serve eighty-two percent of Odisha’s students — had no educational input of any kind for the better part of two years.

The reasons were not mysterious. They were infrastructural. Only twenty-six percent of Odisha’s schools had computers. Only about fifty percent of rural households had smartphones (and many of those were shared among family members, prioritized for adult use). Internet connectivity in tribal and western districts was unreliable or nonexistent. Even where phones and connectivity existed, the content was often inaccessible: parents who were themselves illiterate could not navigate digital learning platforms, download lesson modules, or supervise their children’s engagement with a screen. The DIKSHA platform existed. The content existed. The delivery channel existed in theory. What did not exist was the last-mile infrastructure — device, connectivity, electricity, digital literacy, and an adult facilitator — that would have connected the platform to the child.

The private school experience was different. Private schools in Bhubaneswar, Cuttack, and district towns pivoted to Zoom classes within weeks. Students from families that could afford private school tuition could, by definition, afford smartphones, internet connections, and a quiet room. The learning loss in private schools was real but modest. The learning loss in government schools was catastrophic.

ASER 2022, the first national learning assessment after the pandemic, documented the devastation. Nationally, the proportion of Class 5 children who could read a Class 2-level text fell from 50.5% in 2018 to 42.8% in 2022. The proportion who could do basic division fell from 27.9% to 25.6%. These are national averages. In states like Odisha, where the pre-pandemic base was already low — seventy-five percent of Class 5 students in rural Odisha could not do basic division in 2018 — the pandemic pushed learning levels into a crisis that some educators compared to an outright reversal of a decade’s progress.

The recovery, when it came, was instructive. ASER 2024 showed Odisha among the strongest recovery states, with a more-than-ten-percentage-point improvement in Class 3 reading levels between 2022 and 2024 — placing it alongside Kerala, Himachal Pradesh, and Uttarakhand as recovery leaders. And the recovery was driven primarily by government schools, not private schools. Government schools in rural India actually outperformed private schools in the pace of bouncing back from COVID learning loss.

What drove the recovery? Not digital platforms. Not AI tutoring. Not edtech apps. The recovery was driven by the return of physical schools — teachers in classrooms, students on benches, chalk on blackboards, mid-day meals in plates. The resumption of the institution. The thing that could not be leapfrogged.

The COVID experiment was, in effect, the largest uncontrolled test of the technology-can-substitute-for-institution hypothesis ever conducted. The result was unambiguous. Where institutions existed (private schools, OAVs, well-resourced urban schools), technology helped maintain continuity. Where institutions were weak or absent (the majority of government schools in rural Odisha), technology did nothing. The multiplier multiplied zero.

---

DIKSHA: The Platform That Reached the School That Wasn’t There

The Digital Infrastructure for Knowledge Sharing — DIKSHA — was launched nationally in 2017. It is, by design, an elegant solution. Teachers access professional development modules. Students access interactive lessons, practice exercises, and assessments. Content can be downloaded for offline learning. The platform is free. It is backed by the NCTE. It is integrated into the School and Mass Education Department’s official workflow in Odisha, coordinated through SCERT’s ICT, DIKSHA & NISHTHA cell.

The platform’s national numbers are large. Millions of teachers registered. Hundreds of millions of content pieces accessed. PM eVIDYA integrated DIKSHA as the primary school education delivery mechanism. SWAYAM, the higher education complement, has offered 11,772 courses with 1.21 crore unique registrations and over 4 crore total enrollments. On paper, digital education infrastructure exists at national scale.

In Odisha, the adoption story is thinner. State-specific SWAYAM and NPTEL usage data is not publicly disaggregated, which itself is telling — if the numbers were impressive, they would be published. The ICT cell under SCERT coordinates digital education activities, but the coordination occurs within the constraints documented above: seventy-five percent of schools without computers, erratic electricity in over two thousand schools, limited broadband outside urban areas, and — perhaps most critically — a teaching workforce with limited digital literacy.

NISHTHA — the National Initiative for School Heads and Teachers’ Holistic Advancement — is the in-service teacher training program delivered through DIKSHA. Nationally, 1.5 lakh teachers have been “empowered with ICT skills” through KITE in Kerala. In Odisha, the numbers are harder to find, which suggests the scale is more modest. The training model — typically one-to-three-day workshops — faces a fundamental pedagogical problem: digital tool proficiency is not a skill that transfers from a three-day workshop to daily classroom practice, especially for teachers who are themselves first-generation smartphone users. A teacher who attends a DIKSHA training in the block office, returns to a school without internet, and has no ongoing support or accountability for using the platform, is not going to integrate digital tools into pedagogy. The training becomes a compliance exercise — a box ticked, a photograph taken, an attendance sheet signed.

The deeper problem with DIKSHA in Odisha is not the platform itself but what it assumes. It assumes electricity. It assumes connectivity. It assumes a device — either a school computer (available in twenty-five percent of schools) or a teacher’s personal smartphone (available, but with data costs borne by the teacher). It assumes teacher capacity to navigate the platform, select appropriate content, and integrate it into classroom instruction. It assumes students who have foundational literacy — you cannot learn from a video if you cannot read the subtitles, and you cannot follow a mathematics animation if you have not mastered the concept two levels below.

Each assumption is an institutional precondition. DIKSHA does not build any of these preconditions. It requires them. The platform is designed for the school that already functions. Delivered to the school that does not, it is an app installed on a phone with a dead battery.

---

What Kerala Built: KITE and the Rs 493.50 Crore Lesson

Kerala Infrastructure and Technology for Education — KITE — offers the sharpest possible contrast to DIKSHA’s adoption pattern in Odisha. The comparison is not about technology. It is about what happens when technology is deployed on top of a functional institution versus into an institutional vacuum.

KITE is a state-owned special purpose company under Kerala’s Department of General Education, transformed from the earlier IT@School Project in August 2017. The scale of what it built: 45,000 classrooms in 4,775 schools made “Hi-Tech,” funded by KIIFB (Kerala Infrastructure Investment Fund Board) at a cost of Rs 493.50 crore. Each Hi-Tech classroom includes a laptop, a ceiling-mounted multimedia projector, HDMI cables, a whiteboard or projection screen, USB speakers, high-speed broadband internet, and access to the Samagra Resource Portal — a comprehensive digital content library with e-resources, textbooks, question banks, videos, animations, interactive narrations, and tutorials for Classes 1 through 12.

The numbers are impressive. But they are not the point. The point is what existed before KITE arrived.

Before the smartboard was installed, the Kerala classroom already had: a qualified teacher earning Rs 35,600-75,400 per month (competitive enough to attract genuinely talented people into teaching), consistent electricity, functional toilets, drinking water, a library, and students whose parents had a ninety-four percent literacy rate and who came from a culture where education had been valued as a tool of social liberation for two centuries. The school had been inspected by panchayat-level committees with real authority over teacher accountability. The teacher had attended ongoing professional development delivered by a training infrastructure that invested meaningfully in pedagogy. The school’s performance was monitored through decentralized governance structures that Kerala had been building since the People’s Plan Campaign of 1996, when thirty-five to forty percent of the state plan budget was devolved to local governments.

Into this environment — the functional institution, the trained teacher, the literate community, the accountable governance structure — KITE placed the smartboard. And the smartboard worked. Not because the technology was superior to what other states could purchase (it was the same hardware). But because the institutional environment was superior. The technology amplified what was already there: a competent teacher became more vivid, a good curriculum became more accessible, a strong school became more engaging. KITE also trained 1.5 lakh teachers across Standards 1 through 12 in ICT skills, and trained 52,150 students from 1,990 high schools in animation, electronics, hardware, cyber safety, Malayalam computing, web TV, mobile app development, and robotics. The training stuck because the institutional capacity to absorb it existed.

The VICTERS educational TV channel, also operated by KITE, complemented classroom instruction. During COVID, VICTERS became a primary content delivery mechanism, and it worked better in Kerala than comparable platforms worked in other states — because the students and families receiving the content were literate enough to engage with it.

Kerala’s investment was not Rs 493.50 crore. That was just the KITE budget. Kerala’s real investment was two centuries of social reform that made education a cultural value, a Communist-era land reform that freed lower-caste families from feudal dependence and allowed them to invest in schooling, a library movement that built over ten thousand public libraries across the state, a decentralized governance system that gave local communities genuine authority over school quality, and teacher compensation that made the profession attractive. The Rs 493.50 crore was the last layer — the technology layer — placed on top of an institutional structure that cost vastly more and took vastly longer to build.

The lesson is precise. KITE did not leapfrog institutional quality. KITE enhanced it. The technology worked because the institution worked. The multiplier multiplied something real.

This is the leapfrog that works: technology enhancing a functional institution. The leapfrog that does not work is the one Odisha attempted during COVID and continues to attempt with DIKSHA in schools where the institution itself is broken: technology substituting for what was never built.

---

The EdTech Mirage

While the government side of digital education struggled with infrastructure constraints, the private sector offered its own promise: that commercial edtech could deliver what the public system could not.

The arc of Indian edtech between 2020 and 2025 is instructive. Byju’s, at its peak in 2022, was valued at $22 billion — the most valuable edtech company in the world. It offered slickly produced video content, interactive exercises, and personalized learning paths. Its sales force blanketed India, including Odisha, selling subscriptions to aspirational parents on EMI plans — sometimes to families who could not afford them, sometimes using high-pressure tactics that became the subject of consumer complaints across states. The company expanded into physical coaching centers, acquisitions, and international markets. In January 2024, Byju’s entered bankruptcy proceedings, its valuation collapsed to effectively zero, its financials in disarray, its founder facing regulatory investigations.

The Byju’s trajectory is not just a corporate failure story. It reveals the structural limitation of the edtech model when applied to the populations that need it most. Byju’s primary market was English-medium, urban, middle-class students preparing for competitive exams — JEE, NEET, UPSC, and state entrance tests. The product was designed for students who already had: a quiet room to study in, reliable internet, a personal device, English language proficiency, and a family that could afford Rs 30,000-100,000 per year in subscription fees. The typical rural Odia-medium student — the one sitting on a concrete floor in a school without electricity — was not the target customer. He was not the customer at all.

PhysicsWallah emerged from the Byju’s wreckage as the sector’s most credible survivor. It raised $210 million in 2024 and expanded aggressively, including into offline centers. PhysicsWallah announced content in nine Indian languages, including Odia, signaling an attempt to serve non-English-medium students. The company’s origin story — Alakh Pandey teaching physics on YouTube from his room in Allahabad, building an audience of millions before raising institutional capital — is compelling. It represents the democratization possibility of digital education: a great teacher reaching students who would never have had access to one.

But even PhysicsWallah’s model has structural limits when applied to Odisha’s hardest problem. The company’s core offering is competitive exam preparation. Its expansion into Odia-language content serves students already motivated enough to seek out supplementary instruction, already literate enough to follow structured lessons, and already connected enough to access a digital platform. The student who needs the most help — the one in Class 5 who cannot do division, the one in a tribal district where only twenty percent of students can read Class 2-level text — is not searching for competitive exam preparation. He is struggling with foundational literacy that should have been established three years earlier.

The edtech model, whether Byju’s or PhysicsWallah or Unacademy (which shifted to affordable short-term certification programs after its own contraction), operates on one side of a divide that it cannot bridge. It serves students who are already in the game — those with enough foundation, motivation, connectivity, and resources to seek out additional learning. The students who are out of the game — the eighty-plus percent in government schools who received no instruction during COVID, the sixty-one thousand children the CAG found entirely outside the formal education system — are invisible to the edtech market because they are not customers. They are not paying subscribers. They do not show up in app download statistics. They need something edtech cannot sell: a functioning school.

---

AI Tutoring: The Promise at the Edge

If commercial edtech serves the already-served, artificial intelligence in education offers a different promise: personalization at scale for the underserved.

A 2023 J-PAL study in rural Rajasthan demonstrated measurable learning improvement when personalized, technology-supported learning was introduced in government schools. The key finding was not that AI replaced the teacher. It was that AI-assisted programs allowed adaptation to individual learning levels — particularly valuable in the multi-grade rural classrooms that characterize much of Odisha, where a single teacher manages five grades simultaneously and cannot possibly tailor instruction to each student’s level. An AI tutor that assesses where the student is and delivers content at precisely that level — not two grades above, where the curriculum says the student should be, but where the student actually is — could address the foundational learning crisis in a way that standardized content delivery cannot.

The potential is specifically relevant to Odisha for two reasons. First, the state’s twenty-one-language “Nua Arunima” curriculum for Anganwadi children demonstrates institutional willingness to work in regional languages, including tribal languages. If AI tutoring could operate in Odia and in languages like Kui, Sora, Mundari, and Santali, it could address the triple language barrier — home language to Odia to English — that drives tribal student dropout. Second, the chronic teacher vacancy in remote schools creates a structural gap that AI could partially fill: not replacing the absent teacher, but providing structured learning activities that students can engage with during the hours when no teacher is present.

But the barriers are not trivial. They are the same barriers that limited every previous digital education initiative:

Electricity. 2,182 schools operate without it. AI tutoring software requires a charged device.

Connectivity. Tribal and western districts have unreliable internet. AI platforms that require server-side processing cannot function offline without significant engineering investment.

Devices. Twenty-five percent of schools have computers. AI tutoring at the individual student level requires one device per student, or at minimum one per two or three students. In a school of forty-seven students with zero computers, the arithmetic does not work.

Language models. The AI systems currently available are predominantly trained in English. Odia-language AI is emerging but limited. Tribal language AI is essentially nonexistent. Building language models for Kui or Sora — languages spoken by hundreds of thousands but with minimal digital text corpora — is a research challenge, not a deployment challenge.

Facilitation. Even the best AI tutor requires a human facilitator — someone who manages the devices, troubleshoots technical issues, motivates students, answers questions the AI cannot, and provides the human relationship that is one of education’s three irreplaceable elements. That facilitator needs to be trained, present, and supported. In the school where the single teacher has been absent for a week attending block-level training, the AI tutor sits dark.

The realistic assessment is this: AI-assisted learning has genuine potential to improve outcomes in Odisha’s government schools, particularly for foundational literacy and numeracy in early grades. It could be transformative if three conditions are met. First, the infrastructure prerequisites — electricity, connectivity, devices — are provided. Second, the language technology is developed for Odia and major tribal languages, which requires public investment because the commercial market will not fund it. Third, the AI is deployed as a complement to a functioning school, not as a substitute for one. Give an AI tutor to a school with a trained teacher, functional infrastructure, and community support, and outcomes will likely improve. Give the same AI tutor to a school with no teacher, no electricity, and no facilitator, and you have placed an expensive piece of technology in a room where it cannot function.

The confidence level on AI’s potential is moderate — perhaps sixty percent — because the evidence base from developing-country contexts is still thin, the technology is evolving rapidly, and the infrastructure barriers in Odisha are severe. What is not in question is the direction: AI as complement to institutional quality, never as substitute.

---

Mo School Abhiyan: Building the Institution Through Memory

If DIKSHA and edtech represent the technology-first approach, and KITE represents the institution-first approach, Mo School Abhiyan represents something different: using emotional connection to build the institution from outside the state apparatus.

Launched on November 14, 2017 — Children’s Day — by then-Chief Minister Naveen Patnaik, Mo School leveraged a simple insight: adults are emotionally attached to the schools where they grew up. That attachment could be converted into investment. The programme connected alumni with their alma maters through a digital platform, encouraged financial contributions, and offered a government match at two-to-one — every rupee donated by alumni was matched by two rupees from the state.

The scale was remarkable. Over thirty lakh alumni and community members connected. 50,263 schools transformed through the School Adoption Programme. Alumni contributed Rs 253 crore; the government matched with approximately Rs 506 crore; CSR contributions added Rs 1,025 crore. Total mobilized: over Rs 797 crore across more than forty thousand schools. NITI Aayog recognized it as a best practice in community-driven school transformation. In a single month, 13,302 new alumni were added, contributing Rs 5.33 crore.

The programme was renamed “Panchasakha Sikhya Setu” by the BJP-led government of CM Mohan Charan Majhi in July 2024, with a budget allocation of Rs 332 crore. The renaming is politically routine — incoming governments rebrand predecessors’ schemes — but the programme’s institutional momentum survived the transition, which is itself a sign of genuine value.

What Mo School built was not digital infrastructure. It built physical infrastructure: new classrooms, renovated buildings, functional toilets, compound walls, playgrounds, libraries, computer labs. It built these using a funding mechanism that did not depend entirely on government allocation — the chronic bottleneck documented by the CAG, which found Rs 11,000 crore in education funds surrendered over five years and utilization rates of only forty-four to fifty percent for Samagra Shiksha. Alumni money arrived because someone cared about a specific school, not because a bureaucratic pipeline eventually released funds in March.

The programme also built something less tangible but equally important: community investment in schools. An alumnus who has donated money to their old school will pay attention to whether that school improves. They will visit. They will ask questions. They will hold the system accountable in a way that a distant bureaucrat in the state capital cannot. This is a micro-scale version of what Kerala built through panchayat-level education governance: local ownership creating local accountability.

But Mo School’s limitation is structural, and it maps directly onto Odisha’s internal geography of inequality. The schools most likely to have prosperous, connected alumni are schools in coastal districts — Khordha, Cuttack, Puri, Ganjam — where urbanization, migration to professional careers, and IT employment have produced alumni who can afford to donate. The schools least likely to have such alumni are the tribal schools in Malkangiri, Koraput, Nabarangpur, and Rayagada — where parental literacy is thirty-five percent, where the very concept of “alumni network” assumes an educational tradition that is barely one generation deep, and where the graduates who succeeded have often moved so far from the village that the connection has frayed.

Rs 797 crore across fifty thousand schools is approximately Rs 1.6 lakh per school. For some schools, this funded a computer lab. For others, it barely covered a compound wall. The distribution of investment almost certainly mirrors the distribution of alumni wealth — concentrated in coastal schools, thin in tribal ones. The programme that was designed to transform schools risks reinforcing the very divide it aimed to narrow. The schools that most need transformation are least likely to attract it through alumni philanthropy.

This is not a criticism of Mo School’s design, which was genuinely innovative. It is a recognition that community-driven funding, like technology-driven delivery, works best where the base conditions already exist: an educated community with resources, a school with enough institutional memory to maintain alumni connections, and a location accessible enough that visitors actually return. It is another multiplier. It multiplies what is already there.

---

5T: The Infrastructure Proof

If Mo School built institutions through alumni memory, the 5T High School Transformation Programme built them through state capacity — the most ambitious infrastructure investment in Odisha’s education history.

Launched in 2021, the programme identified 8,800 schools for transformation. The initiative equipped schools with smart classrooms, e-libraries, sports facilities, modern science labs, and renovated campuses. Every participating school received at least two smart classrooms with computers, keyboards, projectors, screens, and televisions. Ninety-six percent of schools received internet connections. All teachers received training on smart board usage.

The impact on national rankings was dramatic. Odisha climbed from rank 24 to rank 5 in the Performance Grading Index 2.0 for 2023-24, surpassing Kerala with a score of 595.6 and achieving the highest grade in the “access” domain. The ranking jump — nineteen positions — was among the largest improvements any Indian state had registered on any education metric.

The enrollment data told its own story. Private school enrollment in Odisha declined from 16.05 lakh in 2019-20 to 14.62 lakh in 2021-22. Students moved from private to government schools — a reversal of the national trend, where government schools were hemorrhaging enrollment to private ones. The implication was clear: when government schools visibly improved, parents responded. The demand for quality education had always existed. What had not existed was the supply.

The 5T programme also produced something harder to measure but equally real: the perception shift. A government school with a smart classroom, a painted facade, functional toilets, and a compound wall no longer looked like a “school for the poor.” It looked like a school. The cosmetic dimension mattered because parent perception drives enrollment, and enrollment drives the political pressure that sustains funding. The virtuous cycle that Tamil Nadu achieved decades ago — visible school quality attracting students, higher enrollment justifying investment, investment producing visible improvement — was beginning to operate in Odisha.

But the critical assessment must be honest about what infrastructure does and does not change. The PGI measures inputs and processes more than outcomes. Odisha’s ranking jump reflected reasonable budget allocation, infrastructure deployment, and policy frameworks — not necessarily learning quality. The PGI anomaly is instructive: Tamil Nadu ranked lower than Odisha on PGI despite having significantly better education outcomes by every learning assessment measure. The index measured the school’s equipment. It did not measure what students learned.

The dropout rate tells the more uncomfortable story. Despite 5T, despite smart classrooms, despite internet connections, Odisha’s school dropout rate rose from twelve percent to fifteen percent in 2024-25. Secondary-level dropout — the point at which students leave school — reached approximately fifteen percent overall, with boys at 17.3% and girls at 12.5%. These are among the highest figures in India. Smart classrooms did not solve the economic pressures that pull boys into labor markets, the migration dynamics that drag families to brick kilns for six months a year, the academic gaps that leave students unable to follow instruction three grades above their actual level, or the absence of local employment that makes continued schooling seem irrelevant.

The 5T programme proved that infrastructure investment can be executed at scale — a non-trivial achievement in a state where the CAG found forty-four to fifty percent fund utilization rates. It proved that visible improvement changes parent behavior. It proved that government schools can compete with private ones when they look and feel competitive. What it has not yet proved is that infrastructure translates into learning. That translation requires something that Rs 42,565 crore in annual education budget has not yet purchased: a teaching workforce trained in pedagogy, not just content; accountable through community governance, not just bureaucratic reporting; and supported by a system that measures what children learn, not just whether schools have projectors.

Technology without pedagogy is furniture.

---

The OAV Model: What Happens When You Build the Institution Right

The Odisha Adarsha Vidyalaya Sangathan operates the closest thing Odisha has to a proof of concept for institutional quality in government education. The model is deliberate in its design: one residential school per block, competitive entrance examination, CBSE curriculum, English medium instruction, dedicated full-time faculty, hostels, laboratories, libraries, and a governance structure separate from the regular government school system.

The OAVs represent what happens when the state decides to build the institution first and let everything else follow. The teacher is recruited specifically for the OAV. She is not a Siksha Sahayak earning Rs 5,000 per month; she is a qualified, full-time, CBSE-trained instructor. The school has electricity not because it is on the grid and the grid sometimes works, but because the infrastructure was designed to function. The library exists not because a circular mandated it but because the school was built with a library room and a book budget. The computer lab has machines that work because someone is responsible for maintaining them.

The result is that OAV students — drawn from rural backgrounds, selected by examination from government school populations — produce outcomes comparable to urban private schools. These are not privileged children. They are the children of farmers, laborers, and small shopkeepers who happen to have performed well on an entrance test. The institution did what the technology could not: it created the conditions for learning.

But the OAV model’s limitation is precisely what makes it work. It is a selected institution serving a selected population. The 315-plus OAVs serve perhaps fifty to sixty thousand students. Odisha has 76.4 lakh students in its school system. The OAVs reach less than one percent. They are islands — the education equivalent of the OSDMA exception documented across this series, where the state demonstrates extraordinary capacity in one domain while the surrounding system remains broken.

The deeper question is structural. Does creating elite institutions within the government system improve the system, or does it create a two-tier structure that allows the regular system to deteriorate? The Navodaya Vidyalaya experience nationally suggests the latter: model schools coexist with system-wide mediocrity, absorbing the most talented students and often the most motivated teachers from the regular system without improving it. The bright child who would have been the informal peer tutor in her village school is now in the OAV. The motivated teacher who would have anchored the local government school is now in the OAV’s faculty. The regular school loses its best participants to the model school. The average declines.

The OAV model proves that the institution matters. It does not prove that technology can substitute for the institution. Every OAV has smart classrooms. But the smart classrooms work because the institution around them works — not the other way around. Install the same smart classroom in the primary school ten kilometers away where students sit on the floor and the teacher opens DIKSHA three times a year, and the outcome will not be the same. The hardware is identical. The institutional software is not.

---

The Numbers in the Room

The data tells a story that resists comfortable conclusions.

Odisha has 61,565 schools. Of these, 2,182 lack electricity. 23,387 — thirty-eight percent — lack tap water. Only 24.9% have computers, against a national average of 32.4%. Only 27.5% have a playground. In 43% of schools checked by the CAG, students sat on the floor for lack of furniture. 611 schools lacked girls’ toilets, and the number was rising, not falling.

The 5T programme transformed 7,000 of the 8,800 targeted schools with smart classrooms and internet. This is 11.4% of all schools. The remaining 88.6% did not receive transformation.

Secondary dropout stands at fifteen percent. Boys at 17.3%. Girls at 12.5%. These rates are roughly double the national average. Between 1.5 lakh and 5.47 lakh children dropped out annually during the CAG’s five-year audit period. 61,487 children were found entirely outside the formal education system. The secondary-to-higher-secondary transition rate was 70.3% — meaning thirty percent of students who completed Class 10 never entered Class 11.

The education budget rose to Rs 42,565 crore in 2026-27 — 13.7% of total state spending, 3.8% of GSDP. The NEP 2020 target is 6% of GDP. Odisha is at roughly two-thirds of the target. But the spending problem is not primarily the quantum. The CAG found that over Rs 11,000 crore in education funds were surrendered over five years. The Samagra Shiksha utilization rate was forty-four to fifty percent. The money existed. It was not spent. And of the money that was spent, less than one percent went to teacher training. One to three percent went to “quality improvement.” The overwhelming majority went to salaries and construction.

Meanwhile, the internal divide is a chasm. Khordha district has 86.88% literacy; Nabarangpur has 46.43%. The gap — forty percentage points — is equivalent to the gap between developed and least-developed nations globally. In Jajpur, 70.6% of Class 3-5 children could read Class 2-level text. In Malkangiri, fourteen percent could. The distance is not geographic. It is civilizational. The same state government, the same education department, the same budget produces a coastal system that approaches adequacy and a tribal system that the researchers compared to sub-Saharan Africa.

The question these numbers raise is not whether technology can help. It clearly can, in the right conditions. The question is whether technology addresses the binding constraint. And the binding constraint in Odisha’s education system is not the absence of smartboards. It is the absence of functional institutions in the schools where the crisis is most severe — the tribal and western district schools where teachers are absent or untrained, where infrastructure is broken, where students migrate with their families for half the year, and where the community has no tradition of formal education to build upon.

A smart classroom does not prevent dadan migration. A DIKSHA module does not compensate for a teacher who earns Rs 5,000 per month and has no incentive structure tied to student outcomes. An AI tutor does not provide the mid-day meal that is, for many children, the primary reason they come to school. An internet connection does not fill the 1,187 vacant university faculty positions that leave higher education staffed by guest lecturers with no research capacity and no institutional commitment.

---

What the Comparators Reveal

The states that solved the education quality problem did not do it with technology. They did it with institutions. Technology came later — and worked because the institutions were already there.

Kerala did not become India’s most literate state because of KITE’s smart classrooms. Kerala became India’s most literate state because Queen Regent Gouri Parvathi Bai declared in 1817 that the state should pay for all education. Because Sree Narayana Guru made education the weapon against caste. Because the Communist government of 1957 brought private-aided schools under government regulation, ensuring quality standards even in non-government institutions. Because P.N. Panicker’s library movement built over ten thousand libraries. Because decentralization gave panchayats real authority over schools. Because teacher salaries were competitive enough to attract talent. Because education compounded across generations — literate parents creating educated children creating literate parents — for two centuries. KITE’s Rs 493.50 crore was the technology layer placed on this two-century institutional foundation.

Tamil Nadu did not achieve five percent secondary dropout because of digital classrooms. Tamil Nadu achieved it because Kamaraj built a school within one kilometer of every habitation in the 1950s. Because Thiagarayar started the mid-day meal in 1922. Because the Dravidian movement made education a political right, not a charitable offering. Because SIPCOT industrial parks were co-located with educational institutions. Because Anna University standardized engineering education quality across 489 affiliated colleges. Because TNSTC connected every town in the state with public transport, so students could physically reach schools and graduates could physically reach jobs. The digital came decades after the institutional.

Himachal Pradesh achieved near-universal enrollment in mountain terrain — objectively harder geography than coastal Odisha — because teacher accountability mechanisms worked, community involvement in school management was genuine, and political consensus sustained education investment across party lines.

The pattern is not subtle. It is stark. Institutional quality first. Technology second. The technology amplifies the institution. It does not create it. No state in India — not one — achieved education quality through digital delivery to broken schools. Every state that achieved quality built the school, trained the teacher, engaged the community, and then added the screen.

---

The Leapfrog That Actually Works

The leapfrog metaphor needs revision.

Developing countries leapfrogged landline telephony with mobile phones. They leapfrogged branch banking with M-Pesa. In each case, the digital tool replaced a function entirely. The phone replaced voice transmission. The app replaced money transfer. The institution — the telephone exchange, the bank branch — was unnecessary because the function did not require it.

Education is not a function to be replaced. It is a capability to be built. The capability requires a relationship (student-teacher), an environment (peers, library, laboratory), and a process (years of structured development). Technology can enhance each of these. It cannot replace any of them.

The leapfrog that works in education is different from the telecom or banking leapfrog. It is not: skip the institution, deliver the function digitally. It is: build the institution faster using technology as an accelerant.

Consider what this means concretely for Odisha:

The 5T programme and Mo School represent institutional investment — building the physical school, the equipped classroom, the connected campus. This is not leapfrogging. This is the hard, expensive, necessary work of constructing the base. It is what Kerala did with teacher salaries and library movements and panchayat governance, what Tamil Nadu did with mid-day meals and school-within-a-kilometer, compressed into a shorter timeline through greater capital expenditure. Whether it produces the same outcomes depends on whether the investment extends beyond hardware to the institutional software: teacher quality, community governance, learning accountability.

KITE-style technology deployment — placing smart classrooms and trained digital pedagogy into schools that already function — is the leapfrog that works. It takes a functional school and makes it better, faster than the school would have improved without technology. It is the genuine productivity enhancer: the competent teacher with a smartboard reaches deeper into her students’ understanding than the same teacher with chalk alone. The school with a digital content library offers more than the school with thirty books. This is multiplication. One times a smart classroom equals a smart classroom. One hundred times a smart classroom equals a transformed educational experience. Zero times a smart classroom equals zero.

AI tutoring, if deployed correctly, could represent a genuinely new kind of leapfrog — not skipping the institution but bridging its gaps. In a multi-grade classroom where one teacher manages five grades, an AI tutor that meets each student at their actual learning level could provide the individualized attention that the teacher structurally cannot. This is not replacing the teacher. It is giving the teacher a force multiplier. But this requires the teacher to be present, the electricity to be on, the device to be charged, and the AI to speak the child’s language. Each requirement is an institutional precondition.

The sequencing matters: institution first, then technology. Or, more precisely: build the institution and deploy the technology simultaneously, ensuring the technology serves the institution rather than pretending to substitute for it. Odisha’s Rs 42,565 crore education budget, if fully utilized (which it has not been — the CAG’s forty-four to fifty percent utilization rate is itself an institutional failure), would fund both. The 5T programme demonstrated the infrastructure can be deployed. What remains is the institutional depth: the teacher training, the community governance, the learning accountability, the connection to employment that gives education its purpose.

---

Cross-References

The digital education question connects to patterns identified across the SeeUtkal research:

The Long Arc, Chapter 6 documented the JIO revolution — the same exponential digital compression that delivers DIKSHA modules to phones also delivers YouTube tutorials, WhatsApp messages, and UPI payments. The hardware has been upgraded. The institutional software has not. The education system exemplifies this gap: the phone exists, the connectivity exists, the platform exists, but the school — the institution that would use them — has not been upgraded correspondingly.

The Churning Fire, Chapter 6 examined language as consciousness technology — how the words a community uses to describe itself shape what it can become. Digital content in Odia is consciousness infrastructure. When PhysicsWallah produces physics lectures in Odia, it creates the possibility that an Odia-medium student could access world-class instruction in her own language. When AI tutors are built only in English and Hindi, they exclude the very students who need them most. The language question in digital education is not a technical detail. It is a question of whose consciousness the technology serves.

Women’s Odisha, Chapter 7 examined the smartphone as threshold — how gendered digital access determines who benefits from the digital revolution. The smartphone that could be an education tool has gendered reach: in households where a single phone exists, it is more likely to be in the man’s pocket than the woman’s hand. A digital education strategy that assumes household device access assumes equal access. The assumption is false. Girls’ access to educational technology is mediated by the same patriarchal structures that determine their access to everything else.

Urbanization Odisha, Chapter 7 identified digital infrastructure as one of six platform layers a city needs. The ninety-six percent internet connection rate in 5T schools is impressive infrastructure. But infrastructure is one layer. Without the other five — land markets, transport, talent retention, governance, network effects — the digital infrastructure sits in schools that exist in communities where educated graduates have no reason to stay.

The Leaving, Chapter 7 examined the diaspora’s digital engagement. If thirty lakh alumni connected through Mo School can contribute Rs 797 crore, the question extends: can the diaspora contribute not just money but mentoring? Digital platforms that connect an engineer in Bangalore with a Class 10 student in Koraput for weekly tutoring sessions would deploy technology at the relationship layer, not just the content layer. This is speculative but directionally sound.

Environmental Odisha, Chapter 4 documented the heat that stays — Titlagarh at 50.1 degrees Celsius, desertification across 42.49% of Odisha’s land area, the ILO estimate that 5.8% of working hours are lost to heat. Climate impact on schooling is physical: heat days close schools, flood days close schools, cyclone days close schools. A digital education system that can deliver learning even when schools are physically closed has genuine climate-resilience value — but only if the infrastructure (devices, connectivity, language-appropriate content) reaches the students affected by closures. The students most affected by climate disruption are in the same tribal and western districts where digital infrastructure is weakest.

---

The Fundamental Question

Can technology fix what institutions could not?

The evidence from Odisha — from COVID’s natural experiment, from DIKSHA’s thin adoption, from 5T’s infrastructure success paired with rising dropout, from KITE Kerala’s contrast, from the edtech boom-and-bust, from the OAV proof of concept — converges on a single answer.

No. Technology cannot fix what institutions could not, because the problem was never a technology deficit. The problem was, and remains, an institutional deficit.

But the answer has a second clause: technology can dramatically accelerate institutional improvement, if the institution exists to be accelerated. The 5T programme proved that infrastructure can be deployed at scale. Mo School proved that community investment can be mobilized. The OAV model proved that government schools can produce quality outcomes. KITE proved that technology amplifies quality. ASER 2024 proved that government schools, when physically open and functioning, drive learning recovery.

The binding constraint is not the screen. It is what happens between the screen and the student: the teacher, the governance, the community accountability, the connection to livelihood that gives education its meaning. The Rs 42,565 crore education budget is sufficient to address these constraints, if — and this is the political economy question that precedes all technical questions — the state can spend the money it allocates (it currently cannot, per CAG), direct spending toward quality rather than construction (it currently does not — less than one percent goes to teacher training), and sustain investment across election cycles (the Mo School to Panchasakha Sikhya Setu renaming suggests policy continuity is fragile even when programme content survives).

The metaphor from computing holds. When hardware advances exponentially but the software running on it has not been updated, the system exhibits bizarre behavior. Odisha’s education system has upgraded its hardware — smart classrooms, internet connections, DIKSHA platforms, AI promises. But the software — teacher training, community governance, learning accountability, employment linkage — is still running code written for a previous generation. The screen is modern. The institution behind it is not.

The screen and the chalk are not substitutes. The screen enhances the chalk. The chalk requires a hand that knows how to use it — a trained teacher, in a functioning school, supported by a community that values what happens inside. Build that, and the screen multiplies it. Skip that, and the screen is furniture.

The question for Odisha is not whether to invest in digital education. It should. The question is whether to invest in digital education before or after investing in the institution that digital education requires to function. The evidence, from every successful comparator, is unambiguous: institution first, screen second. The screen on a functional school is a revolution. The screen on a broken one is a decoration.

---

Sources

Odisha Adarsha Vidyalaya: OAV Sangathan entrance examinations and CBSE affiliation data; UDISE+ school infrastructure data

COVID learning loss: UNICEF, “How COVID-19 Deepens the Digital Education Divide in India” (2020); ASER 2022 and ASER 2024 national findings; World Bank school closure duration comparisons

DIKSHA and digital platforms: DIKSHA Odisha portal (diksha.gov.in/od/); SCERT Odisha ICT, DIKSHA & NISHTHA cell; SWAYAM portal enrollment data (11,772 courses, 1.21 crore unique registrations)

KITE Kerala: KITE official website (kite.kerala.gov.in); DMEO case study on KITE; KIIFB funding data (Rs 493.50 crore for 45,000 Hi-Tech classrooms); KITE Wikipedia entry with implementation scale; Samagra Resource Portal documentation

Kerala institutional context: Education in Kerala historical trajectory; P.N. Panicker and library movement (10,000+ libraries); Kerala People’s Plan Campaign 1996 (35-40% budget devolution); Kerala teacher salary data (Rs 35,600-75,400 monthly)

EdTech landscape: Inc42, “Indian EdTech Preview 2025” (Byju’s bankruptcy January 2024, PhysicsWallah $210 million raise); PhysicsWallah nine-language content expansion including Odia; Unacademy shift to short-term certification

AI in education: J-PAL 2023 study on personalized learning in rural Rajasthan; India AI, “Revolutionizing Primary Education in Rural India”; MindCraft AI education research (ArXiv)

Mo School Abhiyan / Panchasakha Sikhya Setu: Mo School portal data (50,263 schools, Rs 797 crore, 30 lakh alumni); NITI Aayog best practice recognition; OdishaBytes notification on Panchasakha renaming (July 2024, Rs 332 crore budget)

5T High School Transformation: YourStory, “Odisha 5T High School Transformation” (8,800 schools, smart classrooms, 96% internet); Education Today, “Odisha Ranks 5th in PGI 2.0” (rank jump from 24 to 5, score 595.6); JETIR research paper on 5T transformation; UDISE+ enrollment shift data (private enrollment decline from 16.05 lakh to 14.62 lakh)

School infrastructure: UDISE+ 2023-24 and 2024-25 (61,565 schools, 2,182 without electricity, 23,387 without tap water, 24.9% with computers, 27.5% with playgrounds); CAG Audit Report 2018-23 (43% students on floor, adverse SCR in 57% of higher secondary schools, Rs 11,000 crore surrendered, 44-50% Samagra utilization)

Dropout data: Sambad English and OdishaTV reporting on 15% secondary dropout rate 2024-25; CAG audit (1.5-5.47 lakh annual dropouts, 61,487 out-of-school children, 70.3% secondary-to-HS transition rate)

Learning outcomes: ASER 2018 (75% of Class 5 unable to do division; Malkangiri 14% reading level vs Jajpur 70.6%); ASER 2024 (10+ percentage point improvement in Class 3 reading, government schools leading recovery); NAS 2021 dashboard

Education budget: Kalinga TV, “Odisha Budget 2026-27” (Rs 42,565 crore, 13.7% of state spending, 3.8% of GSDP); PRS India budget analysis; CAG fund utilization findings

District literacy data: Census 2011 (Khordha 86.88%, Nabarangpur 46.43%); Academia.edu and ResearchGate district-level analyses

Comparative data: Kerala 94% literacy, Tamil Nadu 82%, India average 77%; Kerala secondary dropout below 2%, Tamil Nadu below 5%, Odisha 15%; Kerala schools with computers ~90% (Hi-Tech programme), Odisha 24.9%

M-Pesa leapfrog: M-Pesa transaction volume and country expansion data; mobile phone leapfrog literature in development economics

SWAYAM/NPTEL/PM eVIDYA: SWAYAM portal data; PM eVIDYA integration framework; SCERT Odisha ICT coordination

Nua Arunima: Digital Knowledge Repository, “Mother Tongue ECCE Curriculum Nua Arunima” (21 languages for ages 3-6); SAMHATI tribal language programme