For five hours, around 30 workers stood at a worksite in Raichur district of Karnataka, waiting – not for work, but for their attendance to be recorded. Despite repeated attempts, only one worker’s attendance was successfully captured over nearly 45 minutes. Many eventually left without having their presence recorded.This is not a story about a minor technical glitch. It is a glimpse into how a legal right to work can be quietly transformed into something conditional – dependent not on whether a worker is present, but on whether a digital system is able to recognise them.As India moves towards a new employment framework to replace MGNREGA, such experiences are not isolated inconveniences; they are early warnings. Even before any formal transition, a quieter transformation is already underway – through the National Mobile Monitoring System (NMMS).As noted in earlier reporting in The Wire, the NMMS system had already reduced attendance to a rigid technical requirement—often experienced by workers as “two photos or no pay.” What is now emerging is a deeper shift: even when workers meet these conditions, the system itself fails to recognise them.Alongside field investigations, a review of regional media coverage across states and languages indicates that these issues are not isolated, with similar problems being reported from multiple parts of the country.Introduced with the promise of curbing corruption, NMMS replaced manual attendance with real-time, geotagged photographic records. The logic appeared straightforward: more data, tighter verification, and less scope for manipulation. Yet the trajectory of NMMS suggests a different outcome. It has not reduced corruption; it has changed its form—while simultaneously excluding workers. In doing so, it has begun to shift MGNREGA from a guaranteed legal right to a system where access depends on digital validation.This shift is not accidental. It reflects a broader direction in welfare governance, where technology is increasingly used not to expand access, but to regulate and condition it.From Digitising Attendance to Reshaping CorruptionIn its early versions, NMMS was relatively easy to manipulate. Attendance could be marked using irrelevant photographs, images of fields, trees, or loosely assembled groups, without any meaningful verification of whether the individuals recorded were actually present. The system did not eliminate corruption; it simply digitised it.Subsequent versions introduced additional checks – headcounts, alerts, and tighter controls over who could record attendance. Facial authentication was later added, along with Aadhaar-based eKYC and device-level restrictions. Each update was meant to close gaps. But each also made the system more complex – and more fragile.Field investigations in Andhra Pradesh, Telangana, and Karnataka reveal a pattern of repeated authentication failures that prevent workers’ attendance from being recorded even when they are physically present. Reports in local media across the country indicate that similar issues are emerging elsewhere as well.In one instance, a worker who had participated in the programme for over a decade found that his attendance could no longer be recorded after facial authentication was introduced. Although his eKYC status showed as complete, the system failed to match his face with the stored image. For three consecutive days, he was unable to work – not because he was absent, but because the system could not recognise him.This is not an isolated issue. Of the approximately 27.3 crore registered workers under MGNREGA, only about 14.7 crore have completed eKYC – leaving a large section already excluded from accessing work. But the problem does not end there. Even among those who have completed eKYC, workers are experiencing repeated authentication failures that prevent their attendance from being recorded. What emerges, therefore, is a layered form of exclusion—where workers are filtered out first by eligibility requirements, and then again by system failures.A System That Excludes Workers but Allows ManipulationThe system is not only exclusionary; it is also inconsistent in how it operates. Workers who are physically present often find their attendance rejected due to repeated authentication failures, yet these failures are unpredictable and difficult to contest.In one reported instance from Telangana, an Adivasi worker whose attendance repeatedly failed despite being present applied talcum powder to his face; the system then accepted his attendance. Whether incidental or not, the episode reveals something deeper: access to work depended not on whether he had worked, but on how the system interpreted his image.According to newspaper reports from Andhra Pradesh, including a documented case from Markapuram district, workers have faced large-scale failures in attendance mapping even after completing eKYC. Field accounts suggest that around 30–40% of workers were unable to get their attendance properly registered, leading to loss of work despite meeting the system’s requirements.Such failures also appear to reshape the terms on which access to work is negotiated. In one village encountered during field investigations, around 150 workers lost their attendance due to NMMS-related failures. When they approached local officials seeking correction, they were asked to pay a lump sum to have their attendance reinstated – effectively turning a system error into a point of extraction.Separately, field observations also suggest that the NMMS application can accept video recordings of a worker’s face as valid attendance – marking presence even when the worker is not physically present – pointing to vulnerabilities that can be exploited in practice.The result is a system with a striking contradiction at its core: it excludes genuine workers while still leaving room for manipulation. The problem, therefore, is not merely one of technical failure but it is a design that neither secures integrity nor protects access.As Rajesh Veeraraghavan argues in his seminal book Patching Development, such technological interventions often do not eliminate underlying problems but are “patched” in practice through informal adjustments and local discretion.Technology Cannot Substitute for AccountabilityThese patterns point to a deeper issue. Corruption in programmes like MGNREGA has never been simply a problem of insufficient data or weak verification. It is rooted in power relations, administrative incentives, and the absence of effective local accountability.Earlier analyses, including in The Hindu, have pointed to a broader pattern: technological interventions in welfare systems often shift the burden of proof onto workers without strengthening accountability.NMMS, however, focuses its scrutiny on workers, who are the most visible and least powerful actors in the system. It does not strengthen institutions such as social audits, does not improve grievance redressal, and does not alter the incentives of those in positions of authority. The system becomes more data-intensive, but not more accountable.What emerges cannot be called a technological problem because it is a shift in governance itself where surveillance increases without a corresponding strengthening of accountability.Each technological upgrade has followed a similar pattern: early versions enabled crude manipulation; later versions reduced some of these possibilities but increased exclusion; newer versions attempt tighter control, yet continue to be both fallible and bypassable.In a programme designed as a legal right to work, this shift has serious implications. When access to work depends on successful authentication, the nature of the entitlement itself changes from a right guaranteed by law to access mediated by a system.From Legal Right to Digital Permission?The NMMS experiment ultimately shows that corruption cannot be engineered away through applications. Without institutional accountability, technology does not eliminate corruption, but it rearranges it, often in ways that are harder to detect and more damaging for those at the margins.More importantly, it risks altering the very character of welfare. A right that is meant to be universal and enforceable begins to resemble a permission, granted not by law, but by whether a system is able to recognise the worker.Addressing corruption in MGNREGA requires strengthening accountability where it actually operates – at the local level. This means reinforcing social audits, ensuring workers have access to verifiable records, building grievance systems that can correct errors quickly, and providing fallback options when digital systems fail.Technology can support these processes. But when it replaces them, the result is not cleaner governance but exclusion layered over opacity.And if this trajectory continues, what is being reshaped is the meaning of a right itself.Venkateswarlu Kuruva and Chakradhar Buddha are senior researchers with LibTech India, a centre within CORD. The views expressed are personal.