Column
The Algorithmic-Human Manager: AI, Apps, and Workers in the Indian Gig Economy
Who manages gig workers — the platform, or its algorithm? A qualitative study published on arXiv interviews 16 gig workers and 21 stakeholders in India's ride-hailing and delivery sector. The findings surface three structural challenges — opacity, unfair outcomes, and misaligned rewards — with direct implications for ESG due diligence, EU AI Act compliance, and fair algorithm design.
Hello. This is Keito Inoshita from Affectosphere Group.
Who manages a gig worker?
When an Uber delivery driver receives a job assignment, determines their rating, or suddenly finds their account restricted — who made those decisions?
Formally, it is the platform company.
In practice, it is overwhelmingly the algorithm.
Task assignment, performance evaluation, rewards, account suspension — most of these are not decided by a human manager reviewing individual cases. They are generated automatically by algorithmic systems designed into the platform.
A study published on arXiv in June 2026 (Omir Kumar, Krishnan Narayanan; arXiv:2606.19975) investigates this reality through qualitative interviews with 16 gig workers and 21 stakeholders in India’s ride-hailing and delivery sector.
Three takeaways for today
- Algorithmic management improves job access and operational efficiency, but generates three structural challenges: opacity, unfair outcomes, and misaligned rewards.
- The paper proposes an “Algorithmic Human Manager Framework” that integrates technological capability with human oversight to address these challenges.
- The findings have direct implications for gig platform design, HR due diligence, ESG disclosure, and EU AI Act compliance.
① What 16 workers said about algorithmic management
The research draws on qualitative interviews with 16 blue-collar gig workers and 21 stakeholders in India’s location-based gig services — primarily ride-hailing and food or package delivery.
The first finding is that algorithmic management does bring genuine benefits.
Job access has improved. Roles that were previously gated by social networks, local connections, or informal hiring practices can now be entered through an app. Operational efficiency has improved as well — route optimization, dispatch matching, and coordination happen at a speed and scale that manual management cannot achieve.
The three major challenges emerge alongside those benefits.
② Three structural challenges
The three challenges the study identifies are significant for anyone thinking about gig platform design or AI-assisted labor management.
The first is opacity.
Workers have no reliable way to know why a particular job was not assigned to them. Rating algorithms are rarely explained. Account restrictions can arrive without clear notification of cause. The algorithmic system makes decisions that substantially affect workers’ livelihoods, but the logic behind those decisions is not visible to the workers subject to them.
The second is unfair outcomes.
Algorithms optimize for average patterns. But real working conditions include factors workers cannot control: heavy traffic in specific zones, neighborhoods where customer ratings are systematically lower, physically demanding delivery conditions that increase completion times. Workers who face structurally harder conditions may receive lower algorithmic scores despite equivalent effort.
The third is misaligned rewards.
Platforms change compensation structures frequently. Workers who have adjusted their behavior to meet a specific incentive structure may find that the rules change before they realize the expected reward. The gap between anticipated and actual compensation erodes trust in the system.
③ The Algorithmic Human Manager Framework
The paper’s proposed response is what the authors call the “Algorithmic Human Manager Framework.”
The core argument: integrating technological capability — algorithmic efficiency — with human oversight — mechanisms by which human judgment can verify, explain, and contest algorithmic decisions — is not a retreat from automation. It is a design requirement for systems that affect workers at scale.
This is not solely a message for platform companies.
For any enterprise that relies on gig workers within its supply chain or operations, the question of how the algorithmic management systems governing those workers operate is increasingly a business risk and governance question, not just a platform ethics question.
Implications for ESG due diligence and EU AI Act compliance
The regulatory and investment context makes this research timely.
The EU AI Act classifies AI systems used for worker evaluation and task assignment as high-risk AI systems. Transparency, explainability, and the availability of meaningful redress mechanisms are among the requirements that will apply. Gig platform algorithmic management is a textbook example of the use cases this regulation targets.
For ESG due diligence, the three challenges this study identifies — opacity, unfair outcomes, misaligned rewards — map directly onto the risk categories that institutional investors and reporting frameworks are beginning to require disclosure on. Whether a company’s supply chain includes gig workers subject to opaque algorithmic management is an emerging due diligence question.
For HR functions, the relevant question is: “When we use gig platform labor, what do we actually know about how the algorithm managing those workers is designed? Is there transparency? Are there mechanisms for workers to contest outcomes? Is the reward structure stable and fair?”
These are not hypothetical future concerns. They are questions that supply chain auditors and sustainability reporting standards are beginning to ask.
Designing fairness into algorithmic management
The research does not frame algorithmic management as an efficiency-versus-fairness tradeoff.
The suggestion is that well-designed algorithmic management can achieve both — but only if fairness is treated as a design requirement from the beginning, not an afterthought.
That means seriously addressing, at the design stage: what is opaque to workers and why; under what conditions does the algorithm generate systematically unfair outcomes for identifiable subgroups; how do reward structure changes affect trust; and what human oversight mechanisms need to exist.
“The algorithm decided” is not going to be an acceptable answer for much longer — not from regulators, not from investors, and not from workers themselves.
That’s it for today!
References
- Omir Kumar, Krishnan Narayanan (2026). The Algorithmic-Human Manager: AI, Apps, and Workers in the Indian Gig Economy. arXiv preprint arXiv:2606.19975.
* This article was written in part with AI assistance and may contain inaccuracies.