In the 18th general elections for the Lok Sabha, which will shape the landscape of future policy, there is little discussion on the jobs crisis in India or the structural factors behind it. This is not for the lack of pain on the ground; there is no dearth of evidence that in terms of precarity, underemployment, and social inequity, the Indian working class is suffering. In terms of electoral promises, the ruling coalition brings nothing new to the table.
Even from the Opposition, at best, we hear of bromides on unemployment, the bare minimum of welfare, often dismissed as “revdi” culture by reactionaries, Universal Basic Income, and guarantees for paid internships. Is there a need for radical thinking that looks beyond social security and redistribution? And to broaden the question, has automation transformed the problem?
The feverish imagination around artificial intelligence (AI) which infects Silicon Valley thinks so. This spectre of AI-led transformation of work is haunting public discourse and stirring speculation in India. It is important to examine this “transformation” with caution, especially the claims by the industry that AI models will be able to replace cognitive labour by performing human-like abstract reasoning. There is no academic consensus that such capabilities exist in any machine learning model, much less that there is progress towards the deeply contested and fantastical notion of Artificial General Intelligence. AI remains a catch-all market term for machine learning technologies, which make probabilistic decisions and/or generate new patterns by learning from existing patterns in large amounts of data. The magic of AI is the magic of scale, its ability to learn patterns becomes the ability to hide the labour behind producing its training data, creating the illusion of intelligence out of thin air.
Deepening the divide with AI
Given the nature of machine learning, that is, learning how to make new decisions from old patterns, the peril is the acceleration of the past, intensifying inequity both at the level of individuals and communities. AI has not transformed the problem; it has accelerated it.
Fortune-telling and fear-mongering by the AI industry is rampant, which simultaneously makes unsubstantiated predictions of existential threats via non-existent technology, while also mystifying existing systems claiming that only the industry knows enough and hence, should self-regulate. This diverts policymakers’ attention from actual job-related issues connected to automation.
Jobs in the age of AI
First, machine learning enables the development of platforms, the ability to police workers remotely making gig work possible. Second, machine learning use cases also monitor and predict the performance of blue-collar and white-collar workers, often using pseudoscientific benchmarks and technologies like emotion detection (a modern variant of the medieval fraud of physiognomy). Both lead to shrinking labour power in negotiating wages, and in India gig work is not even covered under labour laws. Expanding AI thus reshapes workspaces for the worse, and accelerates exploitation.
Why regulation is not a bad word
Historically, the effects of automation, like information asymmetries, deskilling, obscuring the identity of workers, piece-wage, and alienation, have been well-observed and theorised. This is, nonetheless, not on the agenda of the current policy process. These matters are considered private industry concerns. There is a fatalist tech-determinist mindset prevailing in Indian politics which refuses to entertain that technology can follow policy.
There is a recurring theme in policy documents, where India is positioned as a “test-bed” or “garage” for the world; there is a notion that by “solving for India”, we keep up with emerging AI-based technologies. In this peculiar ecosystem, deployment outpaces deliberation, a trend evident in multiple AI adoption indices. This is exacerbated by the entrenched belief in the Indian political mindset that regulation stifles innovation, a notion disproven by polities such as the EU, China, and the US, with comparatively interventionist and, at times, even progressive regulations regarding AI that don’t harm AI research.
Given that this AI ecosystem and platformisation have been accepted by policymakers with hardly an attempt at contesting their existence or fallout, there is also a serious lack of political will to address the consequences on the economic rights of labour. It is not surprising that even in the elections, there is no talk of changing production practices, that is, future industrial policy, banning use cases of tech which violate the dignity of workers like automated surveillance in workplaces, identifying the so-called intermediaries (platforms) as employers, or legislations on exploitative ghost work which lies behind the production of data or its grey market trade. The imagination around workers’ well-being is shrinking and not keeping pace with the capture of this policy space by the tech industry.
The band-aid called welfare
Social security and welfare are not radical solutions, they are milquetoast bandages which do not address production issues around AI and platforms. Examples of this are the Code on Social Security, 2020, which acknowledges new and emerging forms of work, and the Rajasthan Platform-Based Gig Workers (Registration and Welfare) Act, 2023, which confines itself to social safety, without recognising platforms as employers which would force gig companies under labour laws.
Extant narrow AI technology and its various impacts on labour are not some transhistorical force of nature; it is not a genie out of the bottle. It is created and imposed by industry and governments and can be halted, reimagined, and deployed for other ends by democratic interventions. Social security in platforms is a start but should not stop us from concrete and specific policymaking on the AI industry, data practices, and use cases which will require pushing for a mode of production where labour takes precedaence over capital.
Mali is a PhD scholar and Guha is an assistant professor at the Ashank Desai Centre for Policy Studies at IIT Bombay. They work on AI, labour, and policy.
