When PM-KISAN’s system excluded 2.4 crore farmers in 2022-23 because their land records did not match their Aadhaar data, no minister took that decision and no officer signed it. There was no human review and no appeal. That is the defining feature of the algorithmic executive: executive power exercised through code, dashboards and automated determination rather than through Cabinet memos and file notings.
How We Know It Is Happening
PRAGATI. Since 2015 the Prime Minister has chaired monthly video conferences with secretaries and Chief Ministers using real-time data dashboards, tracking over 340 projects worth about Rs 17 lakh crore. Executive oversight now runs through a data platform alongside the traditional Cabinet committee route.
PM-KISAN. Transfers of Rs 2,000 to about 8.5 crore farmers are determined by an Aadhaar-linked land records algorithm. In 2022-23 about 2.4 crore beneficiaries were excluded automatically on data mismatches, without human review or appeal.
CoWIN. Vaccine slot allocation was governed by algorithmic queuing. It enabled 220 crore doses and simultaneously excluded elderly and rural populations without smartphones from early access.
CERT-In directions, April 2022. Mandatory cybersecurity incident reporting rules issued by executive circular, with no parliamentary legislation, imposing a six-hour reporting window on platforms including VPN providers.
Facial recognition policing. Delhi and Uttar Pradesh police deployed facial recognition at protests without any statutory framework or data protection safeguards. The Supreme Court has flagged that such systems constitute State action under Article 12 and are challengeable for arbitrariness under Article 14.
DPDP Act 2023. Government and its instrumentalities are broadly exempt from most data protection obligations, so citizens cannot exercise data rights against State-run algorithms.
What It Has Delivered
The benefits are real and should not be minimised.
| System | Documented outcome |
|---|---|
| Direct Benefit Transfer | ~Rs 35 lakh crore transferred 2013-2024; JAM trinity eliminated ghost beneficiaries; savings estimated at ~Rs 2.73 lakh crore (NITI Aayog 2023) |
| PFMS | Real-time expenditure tracking across ministries; reduced float and parking of funds |
| e-Courts and NJDG | Over 6 crore case status updates digitised, reducing information asymmetry for litigants |
| CoWIN | Enabled the world’s largest vaccination drive; 94 percent adult coverage by 2022 |
| CPGRAMS | Over 2.5 crore grievances filed digitally; average resolution down from about 60 days to 21 |
No argument for accountability should pretend these gains are trivial. Removing ghost beneficiaries at that scale is a governance achievement that manual systems failed at for decades.
The Constitutional Problems
The accountability vacuum. When an algorithm denies a ration card, who is responsible? There is no minister, no officer and often no redressal. This defeats audi alteram partem, the right to be heard, not by refusing a hearing but by removing the forum in which one could be sought.
Non-delegation. The executive cannot constitutionally delegate decision-making power to code without enabling legislation. This happens routinely, and the delegation is invisible precisely because it is technical rather than formal.
Article 21 exposure. Automated exclusion from welfare is deprivation of livelihood. No judicial remedy framework exists for algorithmic decisions, so the right is intact and the route to it is not.
Asymmetric data power. The DPDP exemption means the systems making the most consequential determinations about citizens are the ones citizens have least ability to interrogate.
The Structural Point
The usual framing pits efficiency against rights, which is too simple. The interesting feature of algorithmic administration is that it fails silently and at scale.
A corrupt officer denying a benefit produces a grievance, a name and eventually a case. An algorithm denying 2.4 crore benefits produces a number in a dashboard. The error is larger by orders of magnitude and generates less friction, because there is nobody to complain about.
That inverts the usual assumption that scale brings visibility. Here, scale brings invisibility.
The Way Forward
- Mandate a human in the loop for adverse decisions. Automation may determine eligibility; denial affecting a right should require a reviewable human decision.
- Create a statutory appeal route for algorithmic determinations, with defined timelines, so Article 21 has a remedy attached.
- Require enabling legislation before an automated system makes determinations affecting entitlements, which addresses the non-delegation problem directly.
- Publish exclusion data. Every welfare algorithm should report its exclusion rate and the reasons, since a number nobody publishes is a failure nobody fixes.
- Narrow the DPDP government exemption, at least for systems making individual determinations about citizens.
The state has built genuinely impressive machinery. What it has not built is the paperwork that lets a citizen argue with it.
Frequently Asked Questions
What is meant by the algorithmic executive?
The exercise of executive power through code, dashboards and automated decisions rather than through Cabinet memos and file notings. Decisions that once involved an identifiable officer applying discretion are now produced by software applying rules to data, at a scale no officer could match.
What is the PRAGATI platform?
A platform operating since 2015 through which the Prime Minister chairs monthly video conferences with secretaries and Chief Ministers using real-time data dashboards, tracking over 340 projects worth about Rs 17 lakh crore. It concentrates executive oversight in a data-driven review process that operates alongside Cabinet committee structures.
What went wrong with the PM-KISAN algorithm?
Transfers of Rs 2,000 to about 8.5 crore farmers are determined by an Aadhaar-linked land records algorithm. In 2022-23, about 2.4 crore beneficiaries were excluded automatically because of data mismatches, with no human review and no appeal mechanism. The exclusion was a data artefact rather than a decision anyone made.
What are the documented benefits of algorithmic administration?
Direct Benefit Transfer moved about Rs 35 lakh crore between 2013 and 2024, with the JAM trinity eliminating ghost beneficiaries and savings estimated by NITI Aayog at about Rs 2.73 lakh crore. PFMS enabled real-time expenditure tracking. e-Courts and the National Judicial Data Grid digitised over 6 crore case status updates. CoWIN enabled the world’s largest vaccination drive. CPGRAMS reduced average grievance resolution from about 60 days to 21.
What is the accountability vacuum?
When an algorithm denies a ration card or a benefit, there is no minister who decided, no officer who signed, and often no redressal route. That violates the natural justice principle of audi alteram partem, the right to be heard, because there is no forum in which to be heard.
What is the non-delegation problem?
The executive cannot constitutionally delegate decision-making power to private code without enabling legislation, yet automated systems routinely make determinations that affect rights without any statutory basis for the delegation.
How does the DPDP Act 2023 affect this?
The government and its instrumentalities are broadly exempt from most data protection obligations under the Act. That creates an asymmetry: citizens cannot exercise data rights against State-run algorithms, which are precisely the systems making decisions about their entitlements.
What are the Article 21 implications?
Automated exclusion from welfare can amount to deprivation of livelihood, which engages the right to life under Article 21. No judicial remedy framework currently exists for algorithmic decisions, so the right exists without a route to enforce it.
Practice Questions
Prelims MCQs
- The PRAGATI platform tracks approximately how many projects?
(a) About 90
(b) Over 340
(c) Over 1,200
(d) About 40
Answer: (b) More than 340 projects worth roughly Rs 17 lakh crore are tracked through real-time dashboards. - In 2022-23, the number of PM-KISAN beneficiaries automatically excluded due to data mismatches was approximately
(a) 24 lakh
(b) 1.2 crore
(c) 2.4 crore
(d) 8.5 crore
Answer: (c) About 2.4 crore were excluded with no human review and no appeal mechanism. - Estimated savings from Direct Benefit Transfer, per NITI Aayog, are approximately
(a) Rs 27,300 crore
(b) Rs 1.2 lakh crore
(c) Rs 2.73 lakh crore
(d) Rs 35 lakh crore
Answer: (c) Rs 2.73 lakh crore in estimated savings, against roughly Rs 35 lakh crore transferred between 2013 and 2024. - The principle violated when an algorithm denies a benefit without any hearing is
(a) Nemo judex in causa sua
(b) Audi alteram partem
(c) Res judicata
(d) Stare decisis
Answer: (b) Audi alteram partem, the right to be heard, requires a forum; automated denial often provides none. - Under the DPDP Act 2023, government instrumentalities are
(a) Subject to stricter obligations than private entities
(b) Broadly exempt from most data protection obligations
(c) Required to obtain judicial approval for processing
(d) Prohibited from automated decision-making
Answer: (b) The broad exemption creates an asymmetry in citizens' ability to exercise data rights against State systems.
Mains Questions
- Executive power is increasingly exercised through code rather than through files. Examine the constitutional implications. (250 words)
- Algorithmic administration has delivered measurable efficiency gains and created an accountability vacuum. Critically evaluate. (250 words)
- Automated exclusion from welfare engages Article 21. Discuss the remedies required. (150 words)
- Discuss the non-delegation problem raised by government use of automated decision systems. (150 words)
- Suggest a framework for algorithmic accountability in Indian public administration. (250 words)
Tell Google you want more of this.
Add Anantam IAS as a preferred sourceOne tap, and this site shows up more often in your own Top Stories, AI Overviews and AI Mode. Remove it any time.
