By Padmini Sharma
MILAN, Italy, Apr 10 2023 (IPS)
Excessive reliance on algorithmic management has raised concerns regarding its opaque decision-making mechanisms and implication for workers.
In less than a decade, digital platforms have evolved from a niche market to engulf diverse industries and services across the globe, in developed and developing nations alike.
Defined as online mechanisms that enable exchanging goods, services, or information between different actors, these include the likes of Amazon, eBay, Uber, Deliveroo and Airbnb.
In India, both location-dependent jobs like ride-hailing, food delivery and caregiving to location-independent jobs like crowd work have grown due to the high demand for these services in the market, coupled with huge labour reserves comprising both local and migrant labour forces.
As more than 88 per cent of the total employees in India is engaged in the informal economy, some considered the rise in the platform economy to hold significant potential in addressing existing economic and social disparities.
The term ‘platform economy’ encompasses the growing digital platforms, the models of which are gaining significance over other traditional setups as they offer the possibility to save significantly on structural and labour costs, reduce transaction costs and eliminate barriers.
These have constrained labour force participation across disadvantaged groups and ensure a high degree of autonomy for workers to decide about their workload, work portfolio, time and place of work.
Thus, many workers consider these platforms to extend viable opportunities for earning a living, whether at home or abroad. However, despite these advantages, these platforms have raised concerns over deteriorating working conditions.
Pitfalls of algorithmic management
These platforms depend on algorithmic management to mediate labour relations. In practice this means that algorithms manage labour through certain practices like assigning orders to specific workers, optimising delivery routes, calculating income and incentives, and monitoring and evaluating the performances of workers.
Initially, algorithmic management was seen as a positive development for workers due to its comparison with previous job experiences. Most workers found it to be less stressful, offering them more autonomy and flexibility and above all the belief that the algorithm is more ‘reliable’ in allocating tasks or calculating their income.
Compared to dealing with humans as managers, dealing with apps was a more rewarding experience in the pre-Covid19 era. Undoubtedly, introducing algorithms has its advantages.
When extracting and using massive real-time data, algorithms can execute faster and make more accurate decisions, therefore enhancing workers’ productivity and efficiency while reducing transaction costs.
The use of algorithmic management is seen to have indirect negative implications on the physical and mental health of the workers, which, to meet the targets, are working 14 to 17 hours per day.
Positive as it may seem at first glance, algorithmic management has also introduced certain risks. Although most workers are aware that platforms such as Uber Eats and Deliveroo are strategically leveraging workers’ data to calculate remuneration or assess performances, many workers find it hard to understand the functioning of these apps, in particular the techniques that go into the programming.
This lack of understanding results in doubts about the claimed ‘logical’ and ‘unbiased’ mechanisms of these apps;
‘It does not understand what problems we face on the road […] like when we go to deliver the order to the customer, if there is any problem on the way like a bike accident or anything, then that is not considered […] the company does not understand that […] if I have taken the order, it means I have to deliver it […] and if I am not being able to deliver it, then the app will directly deduct the amount of the order or even its double from the pay-out’, explains a Mumbai delivery worker.
The excessive reliance on algorithmic management has raised concerns regarding these opaque decision-making mechanisms, their implications for workers, their random and inscrutable logic that leaves less room for human comprehension and for workers to contest as well as the high potential for them to propagate existing biases and discrimination.
In addition to this, the use of algorithmic management is also seen to have indirect negative implications on the physical and mental health of the workers, which, to meet the targets, are working 14 to 17 hours per day on average — severely disrupting their work-life balance.
Linking the delivery time to ratings, moreover, makes workers jump traffic signals and ride at high speed, often ignoring the risks associated with such decisions. The assignment of tasks based on several often ‘beyond controllable’ factors by the algorithm increases stress among workers.
These highly controlled unilateral relations with the app are further seen to be disrupting the social relations among the workers which restricts their potential to engage in collective resistance.
Many platform workers are thus moving towards individualistic approaches such as waiting at specific locations or maintaining good terms with the team leaders to make themselves more visible to possibly secure higher orders and income.
Even when some workers are resorting to digital means in uniting, it is not clear whether such mechanisms can contribute towards arousing significant pro-working-class consciousness among the workers.
The challenge of regulating platforms
At the EU level, with multiple cases coming up against algorithmic manipulation and discrimination, and the inaccessibility of data, significant attention is devoted to regulating the rights and interests of platform workers by introducing new governing mechanisms.
As platform workers, with or without support from unions, have brought up several cases against these platforms relating to algorithmic functioning. For example, in Italy, based on the cases filed against app-based delivery platforms, the Courts of Palermo and Courts of Bologna have agreed that the work in these platforms is highly managed via algorithms, the deliveries are assigned based on criteria that are not related to the workers’ preferences or their general interests and that it runs on principles that violate Italian law prohibiting discrimination against employees or self-employed.
The debate in India has mostly centred around including platform workers under the proposed Code on Social Security to ensure more uniform coverage for workers engaged across different platforms.
However, unlike in the European context, the Judiciary in India has not been able to extend recommendations to protect and regulate the interests of the platform or the gig workers. Instead, the debate has mostly centred around including platform workers under the proposed Code on Social Security to ensure more uniform coverage for workers engaged across different platforms.
However, this Code is criticised on several grounds, as it does not solve the main issues concerning workers’ classification and minimum wages and because of its approach to social security, which is still not enough to address existing concerns.
The Code also does not mention any timelines to implement the schemes, thereby adding to the uncertainties of workers. Lastly, the division of powers is also a problem since there is no clear demarcation of responsibilities between the central and state government on labour issues.
A further attempt at regulation in the Motor Vehicles Act of 2020 has sought to place obligations on platforms to maintain transparency over the ‘functioning of the app algorithm’, however, it has not incorporated the ‘right to explanation’, meaning that workers still do not have access to understanding the mechanisms that go into calculating their income, allocating tasks or evaluating their performances.
As workers are coming up with multiple complaints concerning threats to personal data, a lack of transparency, unaccountable algorithmic programming, as well as algorithmic manipulation, there is a strong need to create a more robust governing structure that ensures platform workers greater access to data and to the mechanisms involved in designing their work practices.
Padmini Sharma is a PhD Candidate in Economic Sociology and Labour Studies at the Universita Degli Studi di Milano.
Source: International Politics and Society (IPS), published by the Global and European Policy Unit of the Friedrich-Ebert-Stiftung, Hiroshimastrasse 28, D-10785 Berlin.
IPS UN Bureau