In June 2026, the International Labour Organisation (ILO) adopted the Decent Work in the Platform Economy Convention, 2026. The Convention introduced standards for platform work and prompted broader questions about sectors like data work. As a major hub for outsourcing and services such as data work, India is uniquely positioned to address the labour issues associated with the platform economy and AI value chains. So far, much of India’s AI policy has treated data work primarily as an economic opportunity and enabler for the AI boom, but is yet to explicitly intervene in the sector’s labour issues.The artificial intelligence (AI) sector requires data work (DW) to support the development, deployment, and functioning of AI products. Data workers, by labeling images and providing language data, make important contributions to segments like generative AI, self-driving cars, and robotics. A 2023 World Bank report estimated that the online gig economy hosted between 154 million and 435 million workers. A part of this number is likely to include data workers. Despite its value to AI firms, data work is associated with several labour issues, such as low levels of pay, psychological risksIllustration: Pariplab Chakraborty.India is a major player in the AI boom, being a sizeable consumer market, a seeker of AI solutions, and a veteran outsourcing hub. Data work provides AI firms with data from non-English or “low-resource” languages, enabling AI development and adoption in India. While the Indian state has broached data work, in forms like annotation, these discussions have lacked more direct responses to labour issues and questions of better business practices.India yet to closely examine the sector’s labour issuesWhile lacking a central Act, India has engaged extensively with AI policy and AI actors. Key developments include the 2026 India AI Impact Summit and its resulting Declaration, the 2025 India AI Governance Guidelines (AIGG25), the commencement of the IndiaAI Mission, and 2018’s National Strategy for Artificial Intelligence (NSAI18). The government’s interest in nurturing the AI sector is articulated through undertakings like state-level policies on data centres, the India Semicon Mission, the language-oriented Bhashini platform, and the efforts towards providing compute access to startups. Indian AI policy efforts have covered areas like AI governance, infrastructure, and assisting with data needs.Though not entirely excluded from Indian AI policy efforts, the discussions on data work have been somewhat incomplete. Two notable examples of how data work has been discussed are the AIGG25 and the NSAI18. The AIGG25 introduces the idea of an AI Data Labs Network to develop “grassroots” capacity via instruction in annotating and curating data. In the past, data annotation has also featured in the efforts of the Directorate General of Training (DGT).The older NSAI18 recognises the need for data and developing contextual products while associating data annotation with new kinds of work and responding to the workforce implications of automation. It also discusses a “National AI Marketplace,” which would have a data work component, and mentions a “crowdsourced platform” run by a private actor.The National Education Policy 2020 (NEP20) also discusses data annotation, referring to operations like annotating data. The NEP20 talks about how higher education institutions might provide “ targeted training” for data work to contribute to AI value chains. Instances like the AIGG25 and the NSAI18 indicate that Indian AI policy discussions view data work as important to nurturing the AI sector.If tech firms need data work, then labour should be a part of AI policyFrom the data labs to the data annotation marketplaces, policy discussions lack responses to data work’s labour problems. While Principle 2 of the AIGG25 involves a “people first” approach, its idea of “people” remains narrowly defined. The “people” it centres are primarily AI users and oversight bodies, like those who interact with or govern AI systems. Similarly, the NSAI18’s view also misses discussions of data work’s labour concerns. A more people-first framework would entail engaging with the data work involved and moving past usage.Without measures that assign labour responsibilities to tech firms and AI data solutions vendors, data work could remain an unpredictable and harsh sector that struggles to provide meaningful levels of earnings, upward mobility and safety to workers. Without interventions, data workers could end up shouldering extensive burdens of operating costs, unpaid labour, personal risk, and exacting requirements, without income security, hazard pay and relevant health and safety support. If there exists a state vision of platform-based or outsourcing-driven opportunities providing safety and mobility to people, policy will have to address the data work’s precarious and questionable aspects. Stakeholders need to articulate a template for what data work businesses owe labour, and how the sector fits into people’s longer livelihood trajectories.AI policy’s labour component(s)In addition to promoting sectoral growth and addressing consumer harms, AI policy should guarantee that data workers can rely on a minimum standard of fairness, safety, and viability. Some key issues for AI-labour policy to consider include pay structures, social security and occupational safety, workers’ progression, and a viable working environment.Data work’s task-based and tech-enabled nature may invite comparisons with platform work policy. Beyond the central Code on Social Security, 2020 and the related rules, Indian states have also intervened in the platform economy. For example, Karnataka’s Act has pursued business-facing obligations of grievance redressal, financing social security, and sharing information about automated systems. While such developments have significant potential for data work, they do not constitute a complete solution. For one, data work can involve tech firms engaging labour through vendors across countries. National and state measures may prove insufficient in scrutinising and regulating transnational data work.Data work also requires additional obligations for issues like hazard pay, mental health, support, restricting workers’ voices, and transparency about who and what the work serves. Beyond risks and jurisdictional complexities, data work can involve different models like digital labour platforms and business process outsourcing (BPO). Thus, platform-centric regulation will only partially address the industry’s precarity.Given its imperfect fit with platform work policy, and its role in AI value chains, data work deserves its space in AI policy, alongside topics like synthetic material, intellectual property, and compute (hardware resources that make AI models work). Policy discussions on data work should not be limited to viewing data work as an “AI enabler,” but as a crucial value chain component that needs to meet stringent requirements on the treatment of labour and the handling of data.India’s AI labour policyDespite concerns over labour practices, data work is crucial to the AI sector as both a core value chain component and a market enabler. AI policy needs to find a way to intervene in poor labour practices and pursue healthy, robust data work environments. One essential step involves enabling scrutiny into the sector’s workforce size, labour issues, geographic distribution, and the business relationships between tech firms and data solutions vendors. Evidence-building can inform policy about prevailing problems and barriers while making visible the actors involved.Over time, AI labour policy can create obligations, restrictions, and penalties that stop data workers from falling below basic standards of security and safety. Some important points of intervention include income security, hazard pay, social security, mental health support, and worker collectivisation. For outsourcing arrangements, policy can decide how labour requirements are distributed between entities, potentially promoting better practices across transnational setups.An Indian AI labour policy would mark a major shift in how outsourcing is practised and how it is governed. AI-related labour would be one key issue for the recently formed AI Governance and Economic Group (AIGEG) to consider. Alongside the formation of AI-related bodies, multi-stakeholder or expert bodies on AI policy should include data workers and businesses.Ritvik Gupta and Nighat are researchers with Aapti Institute, a Bengaluru-based research organisation.This piece was first published on The India Cable – a premium newsletter from The Wire – and has been updated and republished here. To subscribe to The India Cable, click here.