Beyond the Employee–Contractor Divide: Platform Work, Algorithmic Control and Labour Protection in India

1. Introduction: the freedom to log in is not necessarily freedom at work

A food-delivery rider may decide when to open an app but have little bargaining power over the delivery fee, the allocation of jobs, the standards used to calculate ratings or the conditions for suspension. The distinction matters. Choosing one's working hours is a form of autonomy; deciding the economic and disciplinary terms of work is another. Conflating the two allows the vocabulary of flexibility to settle a legal question that requires an examination of the actual relationship.

The immediate Indian question has changed since earlier discussions of the gig economy. The Union Government announced that the four Labour Codes, including the Code on Social Security 2020, became effective from 21 November 2025.[note] The Code recognises gig and platform workers and establishes a legislative framework for their social protection.[note] Recognition is consequential, but the legal inquiry cannot stop at a definition. One must ask which persons qualify, what rights follow from that classification, what subordinate rules and schemes govern delivery, and which institutions provide a remedy when a platform or administrator refuses a benefit.

Research question. Can India's recognition and social-security framework protect platform workers adequately without an independent mechanism for examining employment status and algorithmic control?

Argument. No single legal classification is capable of answering every question. A worker may be covered by a dedicated welfare scheme while still asserting employment-based entitlements under another applicable statute. Conversely, an individual may fall outside an employment relationship yet require baseline protection against arbitrary account termination and opaque payment decisions. The appropriate design is layered, rather than a choice between universal employee status and a welfare-only model.

2. Methodology, scope and limits

This is a doctrinal and comparative legal study. It interprets statutory text, reported judgments and official institutional materials, using selected scholarship to identify the policy choices hidden within ostensibly technical definitions. Its principal jurisdiction is India. The United Kingdom is used to test functional status determination; California illustrates the limits of a seemingly bright-line classification test when later exceptions are enacted; and the European Union illustrates regulation of algorithmic management alongside employment status. These jurisdictions are not treated as legally interchangeable.

The manuscript does not undertake interviews, platform data extraction or statistical evaluation of benefit delivery. Assertions about the success of registration drives, actual benefit disbursement or the number of workers covered would require separate administrative or empirical evidence. The legal review is stated as at 28 October 2026, and future amendments or commencement notifications must be checked at proof stage.

3. Why the traditional binary is incomplete

Labour law frequently distinguishes an employee from an independent business operator because the former ordinarily works under another party's direction and lacks equivalent bargaining power. The classification matters for minimum remuneration, leave, working-time limits, collective representation and other rights, although the precise consequences depend on the statute and jurisdiction. Platform work complicates familiar indicators. A rider may own the vehicle, work for more than one app and select working periods; the platform may nevertheless fix the price, define service standards, route customers and remove access to future jobs.

The practical question is therefore not whether a worker has any independence. It is whether the arrangement, assessed as a whole, places material elements of the work under another undertaking's control. De Stefano describes the challenge posed by on-demand work to established protections, while Prassl and Risak analyse how the functions of an employer may be distributed through platform architecture.[note] Their contributions are conceptual scholarship, not rules binding an Indian court.

An analogy helps. A person who rents a shop ordinarily chooses prices, customers and the way business is conducted. A person who is permitted to enter a shop whenever convenient but must accept the owner's prices, customers and discipline exercises a narrower freedom. App-mediated work can resemble the second situation even when the contract calls the individual an independent partner. This analogy identifies the issue; it does not predetermine the legal outcome.

4. India's legislative framework: recognition and its limits

4.1 Statutory categories

Section 2(35) of the Code on Social Security 2020 defines a ‘gig worker’ by reference to work and earnings outside a traditional employer–employee relationship. The Code separately defines ‘platform work’ and ‘platform worker’, including in sections 2(60) and 2(61).[note] The distinction should be maintained: not all task-based work is mediated by a digital platform, and a platform's participation in arranging work does not, without more, answer every legal-status question.

Chapter IX of the Code addresses social security for unorganised, gig and platform workers. Section 114 provides for schemes concerning gig and platform workers and addresses aggregator contributions.[note] The contribution framework is significant because it recognises that the commercial intermediary may bear obligations even where the worker is not treated as a conventional employee. It does not follow that each possible benefit automatically accrues on the date the Code commences. An enabling provision, a notified scheme, a registered beneficiary and a successful claim are different legal and administrative events.

4.2 Commencement is not the same as implementation

The Government's 21 November 2025 announcement establishes a major change in the legal timeline.[note] A subsequent Ministry communication in 2026 addressed the need for governments to frame rules and social-security measures under the Code.[note] The better academic formulation is therefore that the statutory framework has commenced, while individual entitlements and practical delivery depend on the relevant operative rules, notifications, schemes, eligibility conditions and institutional arrangements. This avoids both the obsolete claim that the Code has not commenced and the equally unsound claim that every announced welfare benefit is already universally available.

4.3 Social-security status should not foreclose employment analysis

A statutory definition designed to bring persons within a welfare architecture need not conclusively resolve their status for every other labour-law purpose. Whether a particular claimant is entitled to employment-linked protection must be examined against the governing statutory definition and facts. The danger in treating ‘gig worker’ as an exhaustive label is that the creation of a new protective category could unintentionally be used to deny protections already available to workers who, in substance, satisfy an employment test. A court or regulator should resist that inference unless the legislature clearly requires it.

4.4 State-level experimentation

The Rajasthan Platform Based Gig Workers (Registration and Welfare) Act 2023 offers a state-level registration and welfare model.[note] The Karnataka Platform Based Gig Workers (Social Security and Welfare) Act 2025, Karnataka Act 72 of 2025, provides for a welfare board, registration and specified worker rights, including a grievance mechanism.[note] These enactments are important for two reasons. First, they move beyond general statements of policy towards named institutions and processes. Second, they create questions about coordination: a worker may undertake work across several platforms and move between states, while registration and administration may be organised territorially.

Their significance should not be exaggerated. Enactment of a state law does not by itself demonstrate that its board, database, fund, grievance machinery and benefit schemes are fully functional. Any publication claim about their operation requires the applicable commencement notifications, rules and administrative reports.

5. The comparative case for looking beyond labels

5.1 United Kingdom: Uber BV v Aslam

In Uber BV v Aslam, the UK Supreme Court held that the claimants qualified as ‘workers’ for relevant statutory purposes.[note] The Court emphasised the protective purpose of the legislation and the realities of the relationship, rather than allowing standard-form contracts to determine status. It identified Uber's control over fares and terms, constraints on acceptance and cancellation, monitoring through ratings and restrictions on building an independent customer relationship.[note]

The working-time question was distinct. The Court upheld the tribunal's conclusion, on the facts and applicable regulations, that relevant periods when drivers were logged into the app, in their licensed territory and ready and willing to accept trips counted as working time.[note] It would be inaccurate to turn that holding into a universal proposition that every minute of app availability is compensable under all legal systems.

The case is persuasive for Indian analysis because it shows why contractual form and actual power may diverge. It is not an Indian precedent, and the UK's intermediate statutory ‘worker’ category cannot simply be imported without legislative authority.

5.2 California: the ABC test and its qualifications

In Dynamex Operations West Inc v Superior Court, the California Supreme Court adopted an ABC test in the relevant wage-order context.[note] Broadly, the hiring entity must show freedom from control, work outside its usual course of business and an independently established business. California subsequently enacted Assembly Bill 5, extending the test in important respects.[note]

But an account ending there would misstate the current comparative landscape. Proposition 22 created a separate framework for certain app-based transportation and delivery drivers, and the California Supreme Court addressed its constitutional validity in Castellanos v State of California in 2024.[note] The lesson is not that California provides one settled, universal employee classification. Rather, a strong presumptive test may be substantially modified by sector-specific political and legislative choices. For India, this is a warning against assuming that one formal test will eliminate distributional conflict.

5.3 European Union: status and automated decisions

Directive (EU) 2024/2831 requires Member States to provide effective procedures for determining the correct employment status of persons performing platform work. Article 4 prioritises facts relating to actual performance, including automated monitoring and decision-making, over contractual descriptions. Article 5 provides for a legal presumption where facts indicating direction and control are found, subject to the Directive's qualifications.[note]

The Directive also regulates aspects of algorithmic management, including transparency and human oversight. Its transposition deadline is 2 December 2026; as of 28 October 2026, it must not be described as a fully harmonised domestic employment regime already implemented throughout the Union.[note] Importantly, the Directive does not simply deem all platform workers employees. It requires a legal process attentive to facts and national law.

The earlier manuscript's reliance on Airbnb Ireland is best confined to the separate question of how an online intermediation service is characterised under EU law. That case did not decide whether delivery riders or drivers were employees.[note]

6. Algorithmic management is a labour-law question

Platforms can exercise control without a supervisor issuing verbal orders. Work may be assigned by software; a pricing system may reward particular behaviour; a rating threshold may affect future access; and an automated decision may suspend an account. Rosenblat and Stark's research on Uber drivers is influential in showing how information asymmetries and platform design shape workers' experience of control.[note] Such scholarship supports a research hypothesis, but the precise features of any Indian platform must be proved rather than assumed.

Three issues follow. First, intelligibility: workers should be able to understand the material factors affecting pay, task allocation and disciplinary outcomes. This need not require publication of every proprietary algorithm, but an unexplained adverse decision is difficult to challenge. Second, contestability: a person whose account is suspended or whose earnings are withheld should receive notice, a reasoned route to review and a decision by a competent human reviewer where appropriate. Third, accountability: records must be sufficient to permit regulators and adjudicators to test whether contractual and statutory standards were observed.

The doctrinal rationale is not that every private platform is automatically subject to the full constitutional standards applicable to the State. Rather, these procedural requirements may be created through legislation, enforceable contract terms, sector regulation or specific applicable statutory obligations. The legal basis must be identified in each claim.

7. Social protection, portability and collective voice

Portable protection matters because a worker's livelihood may be distributed across platforms. A scheme that depends entirely on a single platform relationship may leave gaps when the person changes app, works simultaneously for several aggregators or migrates. Registration systems should minimise duplicate records while allowing workers to see contributions and benefits attributed to them. The allocation of financial responsibility should be specified in legislation or scheme rules, not left to voluntary goodwill.

Collective representation presents a different problem. The individual rider rarely has comparable negotiating power over standardised terms. Associations and worker organisations can identify systemic failures in payment, safety and deactivation procedures. Whether particular collective-bargaining rights attach to a person depends on the applicable labour-law definition and legal regime. A welfare scheme should not be portrayed as an automatic substitute for collective rights.

The International Labour Organization's research on digital labour platforms supplies a useful global comparative context for decent-work policy, while NITI Aayog's 2022 report offers India-specific projections and recommendations.[note] Projections should be identified as estimates, not observed headcounts in 2026.

8. A layered reform framework for India

Layer one: status determination. Legislators and adjudicators should apply fact-sensitive tests of control, integration, economic dependence and genuine entrepreneurial autonomy under the relevant statute. A contractual declaration of independent status should be evidence, not a conclusive answer. Where legislation creates a presumption, its scope and burden of rebuttal must be specified.

Layer two: a baseline of rights. Regardless of disputed employment status, a carefully drafted platform-work framework could provide minimum safeguards relating to payment transparency, occupational safety, notice of adverse account decisions, accessible grievance resolution and protection against retaliation for asserting legal rights. These are proposed reforms, not a claim that a comprehensive national code currently grants each safeguard.

Layer three: portable social security. Registration, aggregator contributions, eligibility and benefit claims should work across platforms and, where constitutionally and administratively feasible, across states. Scheme design must address double counting, identity errors, delayed transfers and the possibility that the cost of contributions is indirectly shifted to workers.

Layer four: algorithmic accountability. Platforms should provide comprehensible explanations of consequential automated decisions, retain auditable records and ensure meaningful review. Regulators need powers to obtain necessary evidence subject to privacy and legitimate confidentiality safeguards.

Layer five: evaluation. Governments should publish disaggregated information about registrations, contribution collections, claims submitted, claims allowed, rejection reasons, appeal outcomes and time taken to resolve grievances. Without outcome data, the effectiveness of a statute cannot be established by citing its enactment.

9. Objections and responses

One objection is that increased obligations will reduce flexible opportunities or raise prices. The concern deserves empirical study, not dismissal. Regulation can impose compliance costs, and excessively rigid classification may fail to reflect workers who genuinely operate independent businesses. Yet the existence of costs does not establish that current risks should be borne by workers without remedy. Proportionate obligations, thresholds and phased compliance may address legitimate concerns while preserving minimum protections.

A second objection is that social-security recognition makes status litigation unnecessary. It does not. Insurance and welfare benefits answer different questions from minimum remuneration, working-time rights or remedies for unlawful termination. A worker may need both a portable welfare benefit and a legal determination of the relationship in which work was performed.

A third objection is that algorithmic transparency threatens trade secrets. Full public disclosure of source code is not the only regulatory model. A duty to explain material decision criteria, disclose relevant worker-specific data and permit confidential regulatory inspection may offer more proportionate solutions. The adequacy of each safeguard should be evaluated against the decision's consequences and the possibility of meaningful appeal.

10. Conclusion

India's recognition of gig and platform workers is a substantial legislative development, particularly after commencement of the Code on Social Security 2020. But a definition cannot substitute for an enforceable right, and an enabling scheme cannot substitute for evidence that benefits reach eligible people. Comparative experience reinforces two propositions: employment status should be determined from the realities of work rather than contractual labels, and algorithmic control warrants scrutiny even where employment status remains contested. India's strongest response is therefore neither automatic universal employee classification nor welfare-only exceptionalism. It is a layered system of factual status assessment, portable protection, procedural fairness and measurable implementation.