Is Algorithmic Decision a Bone or Bane

The Nobel laureate and one of the key persons in early quantum theory, Niels Bohr, once said, “How wonderful that we have met with a paradox. Now we have some hope of making progress.” It sounds startling but true. Artificial intelligence (AI) is taking over as the manager. The boss hasn’t disappeared yet. However, it has broken into code.

The organisation of work across cities from Bengaluru to Berlin is becoming more and more an organisation not of people but of algorithms. The systems that allocate tasks, establish targets, decide performance, and silently determine who gets to remain employed. It is not the orthodox narrative of automation as a substitute for labour. It is also not as obvious, but something more, and arguably more important. The outcome is startling, but true. AI is taking over the job of managing human workers.

Digital Shifts in India

Such a transformative nature is quite apparent worldwide. India is not an exception to it. No country is showing this change as much as India. The country has had a proving ground for algorithmic management with its huge pool of labour and the most advanced digital infrastructure, including real-time payments and platform ecosystems. The number of employees who open an application every day is in the millions, not only to find a job, but also to be guided, tracked and evaluated by systems they can no longer see. Whatever you want to call it is the gig economy. But that is a misnomer about the change. The resultant phenomenon is a new labour regime; a decentralised but total control, a regime in which authority is located in design but not in a person.

Consider the case of a delivery rider in Kolkata or Mumbai. The app will determine what orders will be displayed on their screen, the time they have to fulfil the orders and the amount they will be earning. Say no too much work and the prospects become parched. Bonus unlocks and delivers quicker than before. Visibility and revenue are determined by customer ratings, which are usually random. No one to give orders, but behaviour is highly choreographed.

The vow is flexibility. Workers are informed that they are partners, they are free to log in and out, that they are free to make their own hours, to be entrepreneurs in their own labour. The fact is more limited. Volatility is often disguised by flexibility; flexibility is encouraged by incentives which reward 24/7 availability. The employee is not bound and is only so to the extent that an algorithm allows it.

Transformation Worldwide

The model is gaining popularity around the world. Time off task is automatically flagged and warehouse workers are monitored down to the second in the US. Ride-hailing drivers have questioned opaque deactivations in Europe, which are executed due to automated systems. In many parts of Southeast Asia, Africa and Latin America alike, this trend is gaining traction: data-mediated labour, metrics and machine learning.

The peculiarity of India is scale and the smoothness of integration of digital systems. It is possible to onboard, verify and assign work to a worker and pay them in one digital loop. It is productive, friction-free, and it is frequently applauded as innovation. But it also gathers power in a way that we are yet to comprehend.

The focus of this change is data. All actions, such as taking up a job, doing it, and dealing with a customer, are captured, analysed and fed into the system. Algorithms are increasingly predictive, more adaptive and more opaque with time. Those decisions that define livelihoods can no longer be attributed to human decision-making. They are the products of models of which even their inventors might be at a loss to give a full explanation.

Change Vs Imbalances

This forms a sharp form of imbalance. Platforms are omniscient, workers scarcely anything. Why was one of the drivers given more profitable tasks as compared to the other? Why are incentives transformed overnight? Why account being suspended? The responses are hidden in secretive systems and not subject to examination.

Opacity is a non-technical issue; rather, it is a political and economic issue. The traditional labour safeguards were constructed on recognisable employers and workplaces. Both are dissolved in algorithmic management. The employer is turned into a platform; the workplace turns into an application; the manager turns into an unseen system. The rights that are contingent on accountability have trouble establishing themselves. They are already being manifested. A sense of disposability, psychological stress and income volatility are prevalent motifs in the testimonies of workers all across the world. Nuance is lost when the performance is scaled down to measures. In the case of automated evaluation, it is hard to appeal. In the case of control being coded, there is no obvious target of resistance.

It would be a mistake to make this a mere exploitation story, though. There are also real efficiencies brought about by the algorithmic management. It is capable of keeping pace with supply and demand in real time, decreasing idle capacity and opening up income opportunities at scale. To businesses, particularly in highly competitive markets, such benefits cannot be overlooked. To the large number of workers, platform work is still better than the other options they have. Whether or not algorithmic management should exist is not the question. It already does. The question is what type of system we want to accept.

The European Union has been on the path to regulations that provide workers with the right to question and comprehend algorithmic decisions. In the world, courts have started doubting the need to categorise gig workers as independent contractors. Policy discussions on the social security of platform workers have been on the increase in India, albeit with fewer specific frameworks. Not gradual adaptation, but a change of mindset is required. Should algorithms be managers, they need to be held to the same accountability that we hold human authority: transparency, accountability and fairness.

Visibility and X Factor

That starts with visibility. Employees ought to be aware of the distribution of work, the way in which performance is evaluated and the reasons behind punishments. It needs redress mechanisms, actual human scrutiny in case of severe consequences of automated decisions. And it insists on new labour regulations that consider the facts of platform-mediated labour.

India is presented with a special opportunity. After developing some of the most sophisticated digital public infrastructure in the world, it can also be at the forefront of establishing ethical practices of how that infrastructure determines work. It can demonstrate that efficiency and dignity can go together, that scale need not be at the price of justice. A more fundamental cultural change is also in progress. Work has been a traditional source of meaning and identity. It is divided into tasks and metrics by algorithmic management, which is mediated by screens. It is not only an economic risk, but a risk that has an existential nature: an optimised, but not recognised, workforce.

The Way Forward

The invisible manager is here to stay. It will merely become more advanced, entrenched and more challenging to dispute. Our mission now is to make sure that its power is not kept as a secret, that it is brought to light, challenged and finally held accountable. The Future of work is not only about what can be done by machines. It concerns with whom they are permitted to dominate and under what conditions.

(Chakraborty is an Associate Professor in Digital Marketing at TAPMI Bengaluru, MAHE. Saha is an Assistant Professor of Marketing, JK Business School, Gurgaon, India)