The Hidden Bosses in India's Gig Economy: AI, Apps and Worker Rights

Platform work has become a defining feature of the digital economy, especially across the Global South. In India, the rise of smartphones, digital payments, Aadhaar-linked identity infrastructure, and app-based services has made gig work a major source of livelihood for millions. For delivery riders, drivers, warehouse staff and home-service workers, the app is no longer just a tool, but the workplace. It decides who gets work, how much they earn, how they are judged and whether they remain visible to the platform at all.

A new report from the Centre for Responsible AI at IIT Madras shows that this system has opened real economic opportunities, particularly in a labor market short on secure jobs. However, it has also created a new form of vulnerability: workers are increasingly managed by algorithms they do not understand and cannot easily challenge.

Access to Work Has Expanded - So Has Dependence

The platform economy lowers the barriers to finding paid work. Recruitment is faster, onboarding is cheaper and workers can often begin earning within hours. In a country where formal employment remains limited and millions rely on unstable informal labor, that matters. Gig work has helped many workers move from unemployment to income. For some, the app has done what traditional labor markets failed to do: connect them quickly to paying customers.

However, access is not the same as security. One of the report's most important insights is that many workers join gig platforms not because they freely prefer flexible work, but because better alternatives are missing. Gig work often functions less as a model of choice than as a fallback option in a stressed labor market.

If workers depend on apps because other options are scarce, then platforms cannot be seen only as marketplaces or tech intermediaries. They are labor institutions and their rules shape income stability, debt, mobility, welfare and bargaining power. This is crucial for India and other developing economies, where platform work may expand faster than labor protections. A job that is easy to enter but hard to contest can still become a trap.

Accountability Has No Face

The report captures a striking contradiction in how workers understand platform control. Workers constantly interact with the app for orders, ratings, pay, incentives and penalties. Yet many still believe that humans, supervisors, managers or some distant "team in Bangalore", are really making the decisions. It reveals how workers experience algorithmic management: not as neutral code, but as power without visibility. The app tells them what to do, but when something goes wrong, the decision-maker disappears.

Workers interviewed in the study reported uncertainty over how work is allocated, why incentives change, why some workers appear favored and why penalties are imposed. Some suspected that workers who stay logged in longer or newly onboarded workers get better treatment. Others felt the company changes rules without consultation, while expecting workers to absorb the consequences.

This is not just a communication problem, but a governance problem. In traditional workplaces, bad management can at least be identified and confronted. In the platform economy, the supervisor may be a digital system, while the company retains the advantages of control without the visibility of command. The result is a workplace where decisions feel deeply consequential but remain difficult to question.

When the Chatbot Becomes the Gatekeeper of Livelihoods

For ordinary issues, digital support can be efficient. The report notes that workers often rely on chat support for cancellations, customer contact and routine operational fixes. In that narrow sense, automated help can work. However, gig workers are not only dealing with routine issues, but also with livelihoods.

When problems involve pay disputes, penalties, incentive failures or deactivation, workers say chatbot-led systems often become a dead end. Complaints go unresolved, explanations are thin or missing, and there is no clear path to a person with authority to reverse a decision.

If an automated system can affect a worker's livelihood, there should be a human review path. The report's proposed "Algorithmic-Human Manager" model rests on exactly this principle: algorithms can handle coordination and scale, but humans must remain responsible for explanation, override and redress.

The report argues that algorithms can coordinate scale, speed and efficiency, but high-stakes decisions must remain subject to human review, explanation and appeal. A system that can block access to income should not be allowed to hide behind automated responses.

Human oversight is not a sentimental add-on to digital efficiency. It is the minimum safeguard against automated injustice. For governments, this suggests a clear regulatory direction: any system that affects pay, suspension, termination or access to work should be explainable, reviewable and time-bound in its grievance process. Without that, the right to work through a platform begins to look dangerously close to a privilege granted by software.

Efficiency Without Fairness Is Becoming a New Form of Precarity

The report does not deny that technology improves operations. Platforms use location tracking, selfie verification, predictive tools, performance metrics and warehouse automation to reduce fraud, optimize routes and improve throughput. In some cases, technology can even support safety, such as automatic log-offs for fatigued drivers. However, the same systems can become harsh when they are designed for efficiency alone.

Workers described rigid targets, unclear deductions, shifting incentives and penalties for delays caused by traffic, weather, flooding or infrastructure failures. In warehouses and dark stores, automated systems may continue to demand the same output even when physical conditions deteriorate. What looks rational on a dashboard can feel punishing on the ground.

The report terms this as "algorithmic cruelty", which means digital systems can hard-code unrealistic assumptions into everyday work, forcing workers to bear risks they did not create and cannot control.

Pay is where this becomes most visible. Some platforms display earnings in real time, yet workers still struggle to understand how payouts are calculated, why rate cards change or why deductions occur. The result is not transparency, but what the report calls income obfuscation - visibility without clarity.

For workers living close to the edge, that uncertainty is not a minor design flaw, but a form of financial instability. For businesses, a platform that optimizes relentlessly without building trust may gain short-term efficiency, but it also invites churn, backlash, strikes, regulation and reputational damage.

The future should not be "algorithms versus people" but algorithms with people - meaning transparent systems, human review for high-stakes decisions, and stronger social protection for workers whose livelihoods increasingly depend on apps.