Rural jobs scheme app gets AI checks, liveness detection to stop fake attendance
In one line: On 23 October 2026, the Centre upgraded the attendance app of its rural jobs scheme with liveness detection, anti-cloning safeguards and AI photo checks, after face authentication alone failed to stop fake attendance.
Published 24 October 2026. Written by Pratidin from the reports linked at the end; every fact checked by a separate review before publishing. How we work
Union Rural Development Minister Shivraj Singh Chouhan launched an upgraded version of the National Mobile Monitoring System (NMMS) app in New Delhi on 23 October 2026, to verify worker attendance under the Viksit Bharat-Guarantee for Rozgar and Ajeevika Mission (Gramin), or VB-G RAM G, the rural employment scheme that replaced MGNREGA from 1 July 2026. Face authentication was added to the app in March 2026 to curb fake attendance, but field checks still found photographs of photographs being uploaded instead of live images, mismatches in worker counts and gender composition, and the same pictures repeated across muster rolls.
The upgrade adds liveness detection, to confirm that a real worker is in front of the camera and not a printed or old photo; anti-cloning and anti-tampering safeguards against altered or duplicate apps; device-level checks; and AI face verification refined for differences across States and regions. An AI photo de-duplication module flags repeated or suspicious photos and automatically withholds the wage list and Fund Transfer Order until checks are done. Attendance can now be marked offline where there is no network, and an authorised exception mechanism covers elderly workers, persons with disabilities and technical failures. Recording attendance does not by itself trigger payment: wages also depend on the electronic measurement book, Direct Benefit Transfer, spot inspections and social audits.
Since 1 July 2026, the scheme has generated nearly 46 crore person-days, of which about 40 crore were recorded through face authentication, across 31.23 lakh worksites, employing more than 3.65 crore workers from 2.89 crore households. Nine States reported 100% face-authenticated attendance. The VB-G RAM G Act, 2025 guarantees 125 days of work a year per rural household, up from 100 under MGNREGA, with costs shared 60:40 between Centre and States (90:10 for north-eastern and Himalayan States). The upgrade follows allegations by the Uttar Pradesh Congress on 15 October of irregularities in the digital attendance system in 11 districts, including fake attendance through cloned apps. The minister said technology checks should not stop eligible workers from getting their entitlements because of connectivity or technical failures.
Prelims facts
- On 23 October 2026, the Rural Development Ministry launched an upgraded NMMS app with liveness detection, anti-cloning safeguards and AI photo de-duplication.
- Flagged photos automatically hold back the wage list and Fund Transfer Order until checks are complete; offline attendance and an exception mechanism were added.
- VB-G RAM G replaced MGNREGA from 1 July 2026 and guarantees 125 days of work per rural household a year, up from 100.
- Since 1 July 2026, nearly 46 crore person-days were generated, about 40 crore recorded through face authentication, across 31.23 lakh worksites.
- VB-G RAM G costs are shared 60:40 between Centre and States, and 90:10 for north-eastern and Himalayan States.
Quick recall
- When was the upgraded NMMS app launched?
- 23 October 2026, by Rural Development Minister Shivraj Singh Chouhan in New Delhi.
- From when did VB-G RAM G replace MGNREGA?
- 1 July 2026.
- Days of work guaranteed under VB-G RAM G?
- 125 days per rural household a year, up from 100 under MGNREGA.
- Centre-State funding ratio under VB-G RAM G?
- 60:40 for general States; 90:10 for north-eastern and Himalayan States.
- When was face authentication added to the NMMS app?
- March 2026.
- What happens when AI de-duplication flags a photo?
- The wage list and Fund Transfer Order are withheld until checks are complete.
- Person-days generated since 1 July 2026?
- Nearly 46 crore, about 40 crore of them through face authentication.
- Who is covered by the exception mechanism?
- Elderly workers, persons with disabilities and cases of technical failure.
Prelims practice question
With reference to the Viksit Bharat-Guarantee for Rozgar and Ajeevika Mission (Gramin) (VB-G RAM G), consider the following statements:
1. It guarantees 125 days of wage employment per rural household in a financial year.
2. Its costs are shared between the Centre and general category States in the ratio 75:25.
3. States may notify a period of up to 60 days in a year, covering peak agricultural seasons, when no works are taken up.
Which of the statements given above are correct?
- 1 and 2 only
- 2 and 3 only
- 1 and 3 only
- 1, 2 and 3
Show answer
Answer: (c) 1 and 3 only. Statement 1 is correct: 125 days, up from 100 under MGNREGA. Statement 2 is incorrect: the ratio is 60:40 for general States and 90:10 for north-eastern and Himalayan States. Statement 3 is correct: States must announce in advance up to 60 days a year, such as sowing and harvest seasons, when no scheme works are undertaken.
Use this in UPSC Mains: previous-year questions
Recurring theme: Technology, transparency and leakages in welfare schemes
- How to use this
Show that the value of e-governance lies in acting on information: NMMS data now automatically triggers holds on suspicious wage payments.
- Face authentication from March 2026 still allowed photos of photos; the October 2026 upgrade added liveness detection and anti-cloning checks.
- AI de-duplication automatically withholds the wage list and Fund Transfer Order for suspicious entries until checks clear.
- About 40 crore of nearly 46 crore person-days under VB-G RAM G were recorded through face authentication.
- How to use this
Cite the upgraded attendance system as a tool an administrator can use against fictitious workers and false muster rolls, with safeguards for genuine workers.
- Liveness detection and AI de-duplication target fake attendance and repeated photos across muster rolls.
- Payment also needs the electronic measurement book, DBT, spot inspections and social audits, matching work done to payments.
- Offline attendance and an exception route protect elderly and disabled workers from wrongful exclusion.
A legal 125-day work guarantee depends on delivery systems that neither leak nor exclude.
Mains practice question
Technology can curb leakages in welfare schemes but can also exclude genuine beneficiaries. Discuss with reference to the upgraded attendance system under VB-G RAM G. (150 words)
Model answer
On 23 October 2026, the Rural Development Ministry upgraded the NMMS attendance app for VB-G RAM G, which replaced MGNREGA from 1 July 2026, after face authentication alone failed to stop fake attendance.
Curbing leakages
- Liveness detection defeats photos of photos; anti-cloning safeguards target fake apps of the kind alleged in 11 Uttar Pradesh districts.
- AI de-duplication automatically withholds wage lists and Fund Transfer Orders for suspicious entries.
- Layered checks: e-measurement book, DBT, spot inspections, social audits.
Risk of exclusion
- Authentication can fail for elderly or disabled workers and at worksites without network.
- Face matching may vary across regions.
Safeguards built in
- Offline attendance, an exception mechanism through the mate, and region-calibrated AI.
- The minister said technology must not deny eligible workers their entitlements.
Way forward
- Publish failure and exception data; time-bound grievance redress; keep social audits central.
Technology should verify work, not decide entitlement on its own.
The basics
Why this matters
Every welfare scheme faces the same problem: making sure money reaches the person who did the work. Rural employment schemes have long fought fake muster rolls and ghost workers. The October 2026 upgrade of the attendance app under VB-G RAM G is a live case study for GS2 questions on e-governance, transparency and the performance of welfare schemes.
From MGNREGA to VB-G RAM G
MGNREGA, enacted in 2005, gave a legal right to 100 days of wage work a year to rural households. The VB-G RAM G Act, 2025 replaced it from 1 July 2026 and raised the guarantee to 125 days.
- 100 days of work per household a year
- Enacted in 2005
- Replaced from 1 July 2026
- 125 days of work per household a year
- Centre-State cost sharing of 60:40 (90:10 for north-eastern and Himalayan States)
- States can pause works for up to 60 days in peak farm seasons
Why face authentication was not enough
The National Mobile Monitoring System app records attendance at worksites with photographs. Face authentication was added in March 2026. But field checks found photographs of photographs, mismatched worker counts and repeated images across muster rolls. In mid-October, the Uttar Pradesh Congress alleged irregularities in the digital attendance system in 11 districts, including fake attendance through cloned apps.
What the upgrade adds
The new layer rests on Liveness detection and AI checks that sit between attendance and payment.
- 1Mark attendanceWorker's face captured on a verified app, with liveness detection; offline if there is no network
- 2AI de-duplicationRepeated or suspicious photos are flagged
- 3Hold if flaggedWage list and Fund Transfer Order are withheld until checks are complete
- 4Measure workElectronic measurement book records work actually done
- 5PayApproved wages go to bank accounts by Direct Benefit Transfer
- 6AuditSpot inspections and social audits check the record
The scale
The exclusion question
Stronger checks can shut out the very people a scheme is meant for: elderly workers whose faces do not match, workers with disabilities, or sites with no network. The upgrade includes offline attendance and an exception mechanism through the worksite mate. The minister said technology checks should not deny eligible workers their entitlements. Whether that holds will depend on how exceptions and grievances are handled, a theme explored in Leakages and exclusion in welfare delivery.
Go deeper
Why it matters for UPSC
GS2 asks about e-governance, transparency and accountability, and the performance of welfare schemes. The 2025 GS4 paper had a case study on MGNREGA mismanagement, including fictitious workers and poorly kept muster rolls.
The core idea
The VB-G RAM G Act, 2025 guarantees 125 days of rural work. Attendance is recorded on the National Mobile Monitoring System app. Fraudsters beat face authentication with photos of photos and cloned apps. The answer is Liveness detection plus AI de-duplication that automatically holds wages for suspicious entries. The risk is the old tension described in Leakages and exclusion in welfare delivery: stopping fraud without shutting out genuine workers.
Numbers and dates to remember
- 1 July 2026: VB-G RAM G replaces MGNREGA.
- 125 days of work guaranteed, up from 100.
- 60:40 Centre-State funding; 90:10 for north-eastern and Himalayan States.
- March 2026: face authentication added to NMMS.
- 23 October 2026: upgraded app launched.
- About 46 crore person-days, 40 crore face-authenticated; 31.23 lakh worksites; 3.65 crore workers.
Where to go next
- VB-G RAM G Act, 2025: the law behind the scheme.
- National Mobile Monitoring System: the app being upgraded.
- Liveness detection: the technology at the heart of the upgrade.
- Leakages and exclusion in welfare delivery: the trade-off every scheme faces.
Go deeper: can technology clean up a jobs guarantee?
The case for stronger checks. Fake attendance diverts public money from real workers. The ministry's own field checks found photos of photos, mismatched headcounts and repeated images even after face authentication was introduced in March 2026. Allegations from Uttar Pradesh in October 2026 of attendance irregularities in 11 districts, including cloned apps logging fake attendance, show how quickly fraud adapts. Liveness detection, anti-cloning safeguards and AI de-duplication target exactly these tricks, and holding the Fund Transfer Order until checks clear stops money before it leaves.
The case for caution. Technology can also exclude. Elderly workers, persons with disabilities and workers at sites without network are the most likely to fail authentication. A legal guarantee of work under the VB-G RAM G Act, 2025 means little if a failed scan delays wages. The ministry has built in offline attendance, an exception route through the worksite mate and region-calibrated face matching, and the minister said technical failure must not deny entitlements.
What else matters. The app is one layer. Payment also depends on the electronic measurement book, Direct Benefit Transfer, spot inspections and social audits. A good design treats the National Mobile Monitoring System as evidence, not as the final judge.
The governance test. The questions to watch are those of Leakages and exclusion in welfare delivery: how many workers fall into the exception route, how fast held wages are released after checks, and whether grievance redress is quick. Publishing such data would let citizens judge whether the system catches fraud without punishing the honest.
VB-G RAM G Act, 2025
The law that replaced MGNREGA and sets the work guarantee.
In one line: The Viksit Bharat-Guarantee for Rozgar and Ajeevika Mission (Gramin) Act, 2025 replaced MGNREGA from 1 July 2026 and guarantees 125 days of wage work a year to every rural household that seeks it.
Key features
- 125 days of unskilled manual work per household a year, up from 100 under MGNREGA.
- A centrally sponsored scheme: costs shared 60:40 between Centre and States, 90:10 for north-eastern and Himalayan States.
- The Centre sets State-wise normative allocations; States bear spending above them.
- States must announce up to 60 days a year, covering peak sowing and harvest, when no works are taken up.
- Unemployment allowance is retained if work is not given within 15 days of a demand.
- Biometric authentication and geospatial planning are built in.
Why it is in the news
Its attendance system has been upgraded to stop fake entries.
Where to go next
National Mobile Monitoring System
The app that records attendance at rural worksites.
In one line: The National Mobile Monitoring System (NMMS) is the Rural Development Ministry's app for recording attendance at rural job worksites with photographs.
How it developed
The app brings transparency to worksite attendance. Face authentication was added in March 2026 to tackle fake attendance and wages claimed in others' names.
The October 2026 upgrade
The upgraded app adds Google Play Integrity checks on the app and device, anti-cloning and anti-tampering safeguards, liveness detection, region-calibrated face verification, offline attendance, an exception mechanism, and AI photo de-duplication that holds wage lists for suspicious entries.
Where to go next
Liveness detection
The technology that tells a live face from a photo.
In one line: Liveness detection is a check in face authentication systems that confirms a real, live person is in front of the camera, not a photograph, screen or mask.
Why it is needed
Basic face matching only asks whether two faces look alike. A clear photo of an enrolled worker can pass that test. Liveness detection adds a second question: is this a live human being right now? Systems may look for cues such as natural movement or depth that a flat image cannot reproduce.
Why it is in the news
Under the rural jobs scheme, field checks found 'photographs of photographs' being uploaded after face authentication began in March 2026. The upgraded app adds liveness detection to stop this.
The trade-off
Stricter checks can reject genuine users in poor light or with weak devices, which is why an exception mechanism matters.
Where to go next
Leakages and exclusion in welfare delivery
The core trade-off in using technology for welfare.
In one line: Welfare delivery faces two opposite errors: leakage, when benefits go to the wrong people, and exclusion, when eligible people are wrongly denied.
Two kinds of error
Tightening checks reduces leakage but can raise exclusion, especially for the elderly, persons with disabilities and people in areas with poor connectivity. Loosening checks does the reverse. Good design tries to lower both, using several layers of verification and quick ways to fix wrong rejections.
Applied to the rural jobs scheme
The upgraded attendance app tightens checks with liveness detection and AI de-duplication. To limit exclusion, it allows offline attendance and an exception mechanism for elderly workers, persons with disabilities and technical problems. Payment also passes through measurement of work, Direct Benefit Transfer and social audits.
Where to go next
Take the 24 October 2026 quiz: 30 Prelims-style questions with answers