Written by Sanya D. Kishwar, Assistant Professor, Jindal Global Law School, O.P. Jindal Global University. and Sakkcham Singh Parmaar, Student, Jindal Global Law School, O.P. Jindal Global University
- Introduction
In the narrative of India’s AI success story, a dissonant counter-chorus emerges where workers are both its enablers and its casualties. Even as India’s self-congratulatory discourse positions the nation as a rising AI superstar, its hidden producers of data, the new industrial proletariat, are forced to toil in digital sweatshops, receiving barely enough to survive on. Their aspirations to join the ranks of “knowledge workers,” as defined in industry discourse, are thwarted when they find themselves employed in menial, demeaning tasks such as labelling and verifying AI content or annotating images. This article argues that AI-driven work creates a class divide between knowledge and digital workers, with data labourers bearing the burden of powering the AI economy.
Although the hidden nature of such work makes it difficult to obtain accurate statistics, an exploratory study of data workers in a rural setting revealed that, at an Indian data processing agency, working conditions scored only 4.5 out of 10 according to Fairwork Foundation criteria. In other words, these workers were operating at the lowest end of the acceptable range. Even researchers such as Kai-Hsin Hung felt that distribution of the value created, data monetisation, labour practices and protective institutions require further research and policy interventions. Yet such interventions remain conspicuously absent from core policy frameworks, as evidenced by critiques of documents such as the 2025 India AI Governance Guidelines (AIGG25) and the recent NITI Aayog reports, which have been noted for treating data work primarily as an economic enabler rather than addressing the sector’s foundational labour challenges.
Beyond minimum wages and safety conditions, work must also be portable, economically viable, and conducive to professional growth. Failure to intervene risks entrenching systemic inequality and widening the divide between knowledge workers and digital/data labourers. Building on this, this article argues that workers’ conditions need to be re-evaluated in light of AI-driven work processes, and that a right to meaningful work should be the centrepiece of India’s labour laws in the coming decade.
- The Rise of the Digital Precariat
India’s gig workforce has grown from 7.7 million workers in FY 2020 – 21 to 12 million workers as of FY 2025, representing a rise of 55 percent in four years, according to the Economic Survey 2025 – 26. While much of this growth is attributable to the food delivery and transportation sectors, a sizeable proportion are engaged in data work, image labelling, transcription, and data annotation for training artificial intelligence systems. According to the World Bank, the number of people working in the online gig economy was estimated between 154 and 435 million workers worldwide in 2023.
Unlike delivery workers, data workers are distributed across various websites and applications, working from home and remaining invisible to both the public and policymakers. This invisibility makes them the “ghost” workers of the current AI economy. India’s data workers are often hired through multiple layers of subcontracting by technology firms that hire them across borders to create products later sold back to Indians. Thus, while India hosts a considerable portion of the global digital labour force, much of it is “virtual,” with little visibility of its operations. Moreover, a significant number of data workers are graduates who have entered this profession through talent traps, underscoring the emergence of a new category of digital precariat in AI work.
- The Myth of Meaningful Work
The job descriptions for online gig / platform jobs marketed to Indians lead prospective workers to believe they promote a workplace culture of freedom and choice and offer flexible hours and work-from-home options. However, as this study’s findings show, such opportunities remain confined to the language of advertisements. As the researchers note, Data work’s task-based and technology-mediated nature invites comparison with platform work policy, but platform-specific regulation only addresses a subset of risks. Here, platform work refers to work organised through digital platforms, which highlights only part of the risks faced by data workers.
Data work entails risks similar to those associated with platform work, including low pay, psychological hazards, and a lack of clarity regarding employment status. Further, task-based payment encourages speed over accuracy, and the governance of work is complicated by algorithmic management. Reports show that “On some platforms, 43% of workers earned less than Rs 10,000 per month and 70% were unable to meet their monthly expenses.” As Shrinath V, a technology and policy consultant, notes, “It’s a governance issue. Invisible code now wields economic authority.”
Further, as Virginia Mantouvalou, a labour law expert, notes, citing the capabilities approach, the notion of the right to work is best understood as a right to employability, which involves the state assuming duties of teaching and training, as well as ensuring that individuals must have opportunities for meaningful work. In this context, employability means more than access to any job; it requires the conditions for developing and sustaining capability. In the context of data work, none of these conditions holds true. First, the work is fragmented into atomised and odd tasks with little opportunity for professional development. Second, in most cases, the workplace is the worker’s home, leaving little scope to develop capabilities.
- The New Class Divide: A Product of Design
The structural challenges in the Indian labour market are not new, but AI-driven production processes exacerbate many of these issues while creating new ones. A UNDP report on Gender Considerations in the Era of Gen Al notes that there is limited awareness among technologists on how to create inclusive datasets, leading to prolonged under-representation and misrepresentation of identity categories. At the same time, the report refers to a data labour force that is often under-trained for these tasks due to a lack of awareness of good practices, resulting in poor quality of datasets, which may adversely affect the accuracy and quality of AI services. This is a typical case of the labour force being used in the production process while being denied the requisite skills to move up the ladder. In other words, AI design choices do not merely reflect inequality; they deepen it and produce the divide described here. At the same time, the report also points out several genders-based causes for concern. It states that “AI skill penetration among men is estimated at 2.78 compared to 1.65 for women, pointing to an access and usage gap rather than a capability gap. The researcher, Maaz Ansari, Co-founder of Oriserve, notes that women, as a result, are pushed to the lower end of the AI economy. The reason for this is time poverty among Indian women who are more likely to undertake unpaid care work and, therefore, have less scope to upskill. Deepali Kelkar, COO, Secutech Automation, notes that for women in India, multiple disadvantages intersect at the same time: job structure, time constraints, access to technology, and how AI systems are designed. In other words, these structural design choices produce adverse outcomes for women and reinforce the broader class divide through gender inequality.
- The Legal Void
The Labour Codes notified in November 2025 contain certain progressive features, including, for the first time, recognition of gig and platform workers. However, as critics note, they contain provisions that decisively shift the law’s centre of gravity away from the working class. For instance, the Code on Social Security (2020) provides for a 1- 2% contribution by aggregators to a social security fund, but does not cover provident fund (PF), pension, and employees’ state insurance (ESI). Most importantly, these Codes continue to rely on the dichotomy between employers and workers; therefore, they do not adequately address the needs of digital workers who are often hired on a contract basis. Thus, in the case of data workers, the employer-employee relationship is often fudged, denying them the benefits that should accrue to them.
India’s AI policy discourse has been strangely silent on the challenges faced by data workers. The 2025 India AI Governance Guidelines (AIGG25) devote only one section to data workers, titled AI Data Labs Network, which will provide capacity building opportunities in annotating and curating data to enable the creation of grassroots data. This section makes no recommendations on addressing the challenges specific to the people in this data network. Instead, the people who comprise this network are not workers in the traditional sense but users who are required to ensure that the data provided to the AI systems is fit for purpose. This framing weakens the article’s central claim by denying data labour its status as labour and obscuring the divide at the centre of the argument.
Without legally binding obligations for technology companies and AI data solution vendors with explicit references to the labour requirements, the data work environment will continue to be unstable and unsafe and provide no guarantee for stable earnings, mobility opportunities, and protection for data workers. In essence, the AIGG25 recommendations fail to respond to the key issues that determine the exploitation risks and poor performance in the data work sector due to the lack of focus on the framework conditions for workers’ rights and obligations.
- Towards a Right to Meaningful Work
A right to meaningful work, grounded in the capabilities approach, requires viewing work as a human capability that safeguards dignity, voice, and status rather than merely an economic transaction, offering a basis to reimagine India’s labour market framework.
A meaningful work capability would encompass income security and fair pay. Data workers must receive compensation commensurate with their contributions to the economy. Regulation of the platform economy and the introduction of appropriate wage compensation norms have become important. The Chinese model offers some pointers on what such a code could look like. Chinese authorities have mandated minimum earnings standards, social insurance contributions, and algorithmic transparency for firms in the gig economy since 2021. The wage slavery of the platform economy must be countered by creating appropriate training opportunities for such workers. Workers must understand that their skills are marketable and that there is an economic upside to acquiring them. The AIGG25 suggestion of an AI Data Labs Network is a step in this direction, but it must go beyond “capacity building” to articulating opportunities for advancement.
Algorithmic transparency and accountability is equally important. There is a need to introduce appropriate mechanisms to ensure that such workers are not exploited by firms that use opaque algorithms to suppress wages, increase productivity, impose penalties, and eject workers from the platform. Disclosure norms and independent audits of such algorithms are minimum standards to consider.
Additionally, many of these data workers perform highly stressful jobs and must be provided with wages commensurate with the level of stress and hazards they face. The state must expand its social security coverage to this category of workers. Moreover, the ability to join unions and bargain collectively remains an integral element in enhancing worker capabilities. The recent Labour Codes dilute this right by expanding its definition to include mass casual leave and imposing stricter penalties, criminalizing the right to strike and weakening collective bargaining power. Steps must be taken to restore it.
- Conclusion: The Democratic Imperative
The future of work is being transformed by artificial intelligence. But what kind of future does it represent? Is it the dawn of a new era of technological feudalism when a digital aristocracy will exploit a technological proletariat, or will it enable a new take-off toward universal prosperity? Such questions are not merely economic but also political and legal. India’s economic performance cannot be measured solely by GDP growth or even the current account surplus. Sustainable development requires stable, secure, and dignified work. With the coming decade set to witness an explosion in the number of workers engaged in the platform or gig economy, a rethinking of India’s approach to its working-age population is long overdue. The new Labour Codes represent a beginning, but as the Wire article shows, much remains to be done. The ability to secure a right to work must evolve into a right to meaningful work.
As Nuna Zekić observes, labour law must evolve to “cope with the challenges of climate change and the need to transition to a post-growth economy while also addressing the imperatives of social justice and the redistribution of wealth and wellbeing through meaningful work.” While the concepts of degrowth and post-capitalist work might seem like utopian goals, especially for the data labourer in rural India, the realities of the Indian economy are not that dissimilar from the conditions of technological feudalism described by Zekić. Degrowth strategies critiques global capitalism’s pursuit of endless growth for increasing inequality and weakening workers’ bargaining power without delivering universal welfare. In other words, the Indian economy is already a feudal economy in which workers’ bargaining power has declined significantly under the continuous expansion of capital.
The new servile mode of production that characterises the emerging AI economy can be seen as yet another iteration of this process. The promise of growth is being used to justify technological feudalism, with all its attendant inequalities. In India, the expansion of the online or platform economy is likely to exacerbate many of these trends, with adverse consequences for the working-age population. The new Labour Codes present an opportunity to intervene in these trends. However, as this article demonstrates, they remain inadequate as they uphold the very same employer-employee dichotomy, do not recognise the specific precarity of data workers, and offer no scope for professional progression. India must move towards a holistic shift from just a right to work to a right to meaningful work. This necessitates amendments to the Code on Social Security to carve out specific rights for data workers; ensuring algorithmic transparency and compensation and; upskilling and career mobility. Without these interventions, India’s AI-driven growth story will result in a digital feudalism wherein the Indian proletariat will be systematically exploited by a tiny elite of technocrats. The future of work in India is not just a matter of democratic governance and social justice; but also one of its basic rights as a working-age population. The time for action is now!
Caveat: The views, analyses, and information presented in this article are provided in good faith and for general informational purposes only. No representation or warranty, express or implied, is made regarding the accuracy, adequacy, validity, reliability, or completeness of the information. Readers should conduct their own research and seek professional guidance where appropriate. Neither the author nor the publisher shall be held responsible for any loss, liability, or consequence arising from reliance on this content.
Leave a Reply