Algorithms vs. Agency: Redefining the Master–Servant Relationship in the Age of the Gig Economy

From Dharangadhara to Dashboards. How India’s New Labour Codes Confront the Algorithmic Black Box. "The law has built a bridge, but placed it two feet above the water. Those who cannot swim are still drowning."

The Illusion of the Log-Off

On New Year’s Eve of 2025, thousands of delivery partners in Bengaluru, Mumbai, and Delhi collaborated in a strike dubbed the “Great Log-Off,” where they declined to take up any orders during the service’s peak period of demand. They did not picket a factory gate. They simply pressed a button. The platforms, in response, activated algorithmic deterrents: lowered acceptance-rate scores, withheld quest bonuses, and nudged substitute workers with surge multipliers. The workers were invisible to any human manager, yet they were being managed, relentlessly, invisibly, and at scale.

This is the central paradox of gig labour in 2026: a workforce marketed as a community of independent entrepreneurs is, in practice, subject to a form of control more pervasive than the most vigilant factory foreman. The law, for most of India’s post-independence history, had a ready answer to the question of who was an “employee”, the Control Test laid down in Dharangadhara Chemical Works Ltd v State of Saurashtra (1957).^[1] A master told the servant what to do and how to do it. In 2026, the “master” does neither, and yet it does both.

The constitutional dimension of the Supreme Court's ongoing PIL in Indian Federation of App-Based Transport Workers (IFAT) v Union of India (2021)^[2] is encapsulated in the following constitutional question: Does exclusion from the category of “employee” amount to a violation of the constitutional rights of gig workers under Articles 14 and 21 of the Constitution? In its current hearing in February 2026, the Hon’ble Supreme Court has specifically asked the Union government about the “subordinate legislation gap,” where, although platform workers have been included as a category under the Code on Social Security, 2020, the schemes for health and accident insurance are still on the drawing board.

The empirical landscape is stark. As the Economic Survey of India 2025-26 ^[3]states, “In India, for FY 2025, the gig workforce size is at 12 million, which is expected to rise to 23.5 million by 2030 with 15 billion UPI transactions and 800 million+ smartphone users.” Yet, “40 percent of these gig workers in India are paid less than ₹15,000 as monthly earnings, excluding fuel and platform charges.”

Further, as revealed in a research article published in the Indian Journal of Integrated Research in Law (2025), ^[4]“Majority of urban gig workers in India are from Scheduled Castes, Scheduled Tribes, and Other Backward Classes, implying that when algorithms are discriminatory, centuries of exclusion are being compounded.”

The Legal Shift: The Code on Social Security and the “Middle Path”

The Code on Social Security, 2020, which came fully into operational effect on 1 April 2026, represents India’s legislative attempt at a “middle path.”^[5] Rather than forcing gig workers into the binary of “employee” or “independent contractor,” the CoSS creates a new statutory category, the Platform Worker, defined as a person who accesses an organisation’s or individual’s application or platform to provide services or do work and whose remuneration is determined by such platform.

The centrepiece of this framework is the Social Security Fund. Aggregators like Zomato, Swiggy, Ola, Uber, Urban Company, etc., are now required to contribute between 1 and 2 percent of their annual turnover to this fund. In return, platform workers are now entitled to accident insurance, health cover, and provident fund contributions. It is, on paper, a historic achievement for a workforce that for a decade operated entirely outside the social security net.

The Eligibility Cliff: When the Safety Net Has a Trapdoor

The celebration, however, needs to be tempered. Draft rules that were proposed in early 2026 indicate that a worker must be active for 90 days on a single platform or 120 days on multiple platforms in a financial year.  The socio-legal consequences of this threshold are profound. Research suggests that up to 60 per cent of seasonal and migrant gig workers may never cross this line. A migrant working as a driver with Ola during the Diwali season and reverting to his job as a labourer in the fields, a homemaker working as a delivery partner with Swiggy on school days alone, a student working as a delivery partner with Rapido on weekends alone, the most economically vulnerable and the least protected are the ones who will be included legally but excluded in practice. This is what scholars have begun to call “digital precariousness.”^[6], a state where the worker is visible enough to be exploited by the platform economy but invisible enough to be denied protection. The CoSS, in this reading, performs inclusion while structuring exclusion.

The Socio-Legal Core: Algorithmic Management as “Control”

The most urgent theoretical challenge the gig economy poses to Indian labour law is conceptual, not merely remedial. The traditional legal architecture rests on the Control Test from Dharangadhara Chemical Works.^[7]A master exercises control over the manner in which work is performed; this was the hallmark of the employment relationship. The independent contractor, by contrast, controls their own method; they are responsible only for the result.

Algorithms shatter this binary. A Zomato delivery partner is told by the application where to pick up the order, which route to take, what delivery window is acceptable (10 minutes), and what consequences will follow from non-compliance, a lowered rating that diminishes future order allocation. No human manager issued a single instruction. Yet the degree of operational control is extraordinary. This is Algorithmic Management: supervision that is automated, continuous, and invisible, operating through what scholars have called the Digital Panopticon.^[8]

The Nudge as Command

The subtlety of algorithmic control lies in its mechanism of nudging rather than ordering. “Quest bonuses” ₹200 extra for completing 10 rides before noon are technically incentives. In practice, for a worker earning ₹15,000 a month, they function as commands. A study cited in the Indonesian Journal of Socio-Legal Studies (2025)^[9] documents the phenomenon of “Time Poverty”: the algorithm rewards workers who are active during social and festival hours, effectively penalising participation in family and community life. The algorithm reaches into the worker’s home.

Perhaps most alarming is the spectre of Algorithmic Wage Discrimination. Platforms possess granular data on each worker’s financial behaviour, withdrawal patterns, frequency of accepting lower-value orders, and historical desperation signals. The theoretical capacity to offer lower pay to workers the algorithm “knows” are more financially desperate has been flagged by researchers as a predictable next frontier of platform exploitation.^[10]

The Persuasive Authority of Uber BV v Aslam

The UK Supreme Court’s landmark ruling in Uber BV v Aslam (2021)^[11] offers the most powerful persuasive precedent available to Indian courts. The Court unanimously held that Uber drivers were “workers” (not independent contractors) because the reality of control over fares, routes, and performance metrics trumped the contractual label of partnership. The principle is directly transposable: Indian courts should look past the “Partner Terms of Service” to the lived architecture of algorithmic management. The substance of the relationship, not its legal costume, must determine its classification.

The Karnataka Model: A Blueprint for the Nation

While the Centre navigated its “middle path,” the State of Karnataka stepped decisively forward. The Karnataka Platform-Based Gig Workers (Social Security and Welfare) Act, 2025^[12] It is the most sophisticated gig labour legislation in the sub-continent. It is the first legislation in India that: Mandates a 14-day notice period before deactivation of a worker’s platform account, as “de-platforming” is equivalent to termination; Creates an Internal Dispute Resolution Committee that has jurisdiction over algorithmic disputes; and confers upon workers the Right to Information regarding the basis upon which ratings and allocations are computed, a right to algorithmic transparency.

This last provision is the most radical. Drawing on the logic of GDPR-style “right to explanation” protections in Europe, the Karnataka Act demands that the Black Box be opened at least partially to the workers whose livelihoods it governs. The corresponding Karnataka Platform-Based Gig Workers Rules, 2025^[13] Further require that any termination initiated by an algorithm be reviewed by a human officer within 48 hours. This “human-in-the-loop” requirement directly confronts what scholars have identified as the accountability vacuum of automated HR management.

The Digital Personal Data Protection Act, 2023 (DPDP Act)^[14] provides a complementary constitutional scaffold. Workers subjected to invasive behavioural tracking of their idle time, their route deviations, and their biometric logins should, under a purposive reading of the DPDP Act, have rights against such surveillance as data principals. The argument that gig platforms constitute “significant data fiduciaries” capable of impacting the fundamental rights of workers is, as of early 2026, yet to be fully tested before the courts, but it is ripe for litigation.

From Welfare to Rights: The Path Forward

Algorithmic Due Process

Algorithmic Due Process is the principle that any automated decision materially affecting a worker’s livelihood, a rating drop, an order suspension, or a deactivation must be explicable, contestable, and reversible. It draws on Article 21’s guarantee of a fair procedure before the deprivation of livelihood,^[15] as well as the Karnataka model’s Internal Dispute Resolution architecture. The demand is not that platforms reveal proprietary source code, but that they provide a worker-intelligible explanation: why your account was flagged, what data was used, and how you may respond. As Pradip Ninan Thomas argues in The Gig Economy in India: Start-Ups, Infrastructure and Resistance (Routledge, 2026),^[16] venture capital logic systematically builds informality into platform design. Algorithmic Due Process would be the counter-pressure of constitutionalism.

Portable Benefits

The second pillar is portability. Under the current CoSS framework, social security entitlements are platform-specific; a worker who moves from Swiggy to Zepto effectively starts again from zero. A truly protective framework would link benefits to the worker’s Unique Account Number (UAN) or e-Shram registration, not to any particular aggregator. As the worker moves across platforms, which is the economic reality for the majority of gig workers, their accumulated entitlements travel with them. This is the architecture of security; anything less is, as the CoSS presently stands, a mirage.

Conclusion

The story of Indian gig labour law is, at its deepest level, a story about who counts as a person in the eyes of the law. The Dharangadhara control test was designed for a world of foremen and factories. The algorithm is its twenty-first century successor, more powerful, less visible, and currently more immune from legal accountability. The Code on Social Security, 2020, marks a genuine legislative advance: for the first time, the 12 million Indians who drive, deliver, and care for strangers through applications have a statutory identity. But identity without enforceable rights is a political gesture, not a legal guarantee.

The path forward runs through three landmarks: the Supreme Court’s eventual ruling in IFAT v Union of India (2021) ^[17], the national replication of the Karnataka model’s transparency and due process provisions, and the extension of DPDP Act protections to the workplace. Until then, the gig worker remains what the sociologists call “independently dependent”: legally free, operationally captive, and algorithmically managed by a master who has no name, no face, and no legal address. The law must catch up to the code.

---

This blog has been authored by Skand Vats and Riyansh Gupta, students at University Institute of Legal Studies, Panjab University, Chandigarh (6th Runner Up of 1st RGNUL National Socio-Legal Blog Writing Competition)

---

REFERENCES

[1] Dharangadhara Chemical Works Ltd v State of Saurashtra AIR 1957 SC 264 (1957)

[2] Indian Federation of App-Based Transport Workers (IFAT) v Union of India Writ Petition (Civil) No 1068 of 2021

[3] “Economic Survey” (Ministry of Finance, Economic Survey of India 2025-26 (Government of India 2026) Ch 5) <https://www.indiabudget.gov.in/economicsurvey/> accessed March 13, 2026

[4] “Socio-Demographic Profile of Urban Gig Workers’ Indian Journal of Integrated Research in Law (2025) 5(2)”

[5] Code on Social Security 2020 (No 36 of 2020) ss 114-116

[6] Constitutional Rights in the Age of Algorithms: Gig Work in India’ Indian Journal of Integrated Research in Law (December 2025)

[7] Supra Note 1

[8] Recalibrating Labour Process Theory for Algorithmic HRM’ International Journal of Innovative Business Strategies (January 2026)

[9] Legal Protections for Gig Workers: A Comparative Socio-Legal Study, Indonesian Journal of Socio-Legal Studies (2025)

[10] Supra Note 4

[11] Uber BV v Aslam UKSC 5, [2021] 4 All ER 209

[12] Karnataka Platform-Based Gig Workers (Social Security and Welfare) Act 2025 (Karnataka Act No 12 of 2025) ss 8, 11, 14.

[13] Karnataka Platform-Based Gig Workers Rules 2025, r 18.

[14]Digital Personal Data Protection Act 2023 (No 22 of 2023) ss 4, 7, 8

[15] Olga Tellis v Bombay Municipal Corporation AIR 1986 SC 180.

[16] Anand Raj, “Book Review: The Gig Economy in India: Start-Ups, Infrastructure and Resistance. By Pradip Ninan Thomas The Gig Economy in India: Start-Ups, Infrastructure and Resistance. By Thomas Pradip Ninan. New York: Routledge, 2025. 144 Pp. ISBN 9781032793740, $144.50 (Hardcover); ISBN 9781032793757, $45 (Paperback).” [2025] ILR Review

[17] Supra Note 2