It’s hard to recall when “the algorithm” wasn’t part of everyday life. Much of that is thanks to digital platforms like Uber, which have made it common to coordinate an entire workplace around lines of code. The International Labour Organization estimates that worldwide, there are now about 800 digital platform companies. From a few core brands in ride-hailing, they’ve multiplied to cover anything from sprawling super apps to lightning-fast delivery and microtasking.

“Algorithmic management” is key to how these platforms work: Workers aren’t managed by human bosses, but are optimally matched by a platform with customers and tasks in the name of efficiency. As platform work has swelled, the downsides of this for workers have become clear. Some pressures are familiar — on earnings, for example. Others are less obvious, like workers being summarily suspended from an app or randomly assigned jobs in dangerous areas.

As part of our landmark report series on digital labor platforms, the Tony Blair Institute for Global Change engaged with 57 platform workers and industry groups in Singapore, London, Nairobi, and Jakarta. (Their identities, as well as those of the platforms they work with, have been anonymized in this article for data privacy purposes.) 

Across the board, these workers still believed that gig work, more so than traditional employment, was likely to be “good work” — something they defined by control, flexibility, and a decent income. But the tensions of algorithmic management were just as clear. We found that relying on automated processes to allocate jobs, rate performance, and determine pay risks undermined those very benefits. And as algorithmic management becomes more common in white-collar work, it’s worth seeing its impact in the gig economy as the canary in the coal mine.

In interviews, gig workers described the lack of transparency about how algorithms work. They want to know what data is collected and how they can respond. One Singaporean worker pleaded, “Just tell me how [algorithmic controls] affect my earnings, so that I can adjust accordingly.” 

That lack of clarity played into confusion over how their performance was being assessed. In Nairobi, workers described accepting jobs in high-risk areas, fearing that rejecting the job may affect their ratings. In Jakarta, they felt that platforms took a “sanctions-first” approach to customer complaints, often without any human intervention. One delivery rider, when suspended from the app, pointed out that the action wasn’t accompanied by “a chat or email for us to clarify [why].” 

As an extension of that, commissions taken by the platform can increase without consultation, and the calculations behind payment rates can be unclear. One ride-hailing worker in Indonesia reported that, confusingly, rates or fares may not match the actual distance traveled. “Fares for 3 kilometers can be the same as 1 kilometer,” he said. 

We also found that price discrimination may exist, based on workers’ locations. Online workers from Jakarta and Nairobi found that customers “tend to think we do not know or understand English — hence, we get fewer tasks and low-paying jobs.” Some web-based platforms can also restrict access to markets by location; one transcriber in Kenya reported having to use a VPN.

Why is this important? Algorithmic management isn’t confined to emerging markets or to digital labor platforms. As of July 2020, 42% of European firms had adopted at least one algorithm-powered technology, while a further 18% were considering implementing one in the next two years. Today, tech-enabled tracking and surveillance tools are deployed in retail, warehouse, and logistics businesses to collect data about the speed, behavior, and compliance of workers. These practices can have a profound impact on workers’ well-being, income, sense of fairness, and autonomy.

We need to preserve the benefits of gig work while mitigating its harmful aspects, for the good of the whole workforce. Some governments have chosen to implement protections and benefits that are tied to employment status, reclassifying workers as employees. This sets up a false choice between flexibility and protections for workers. 

In contrast, other governments, like that of India, have delinked worker protections from worker status, which accords greater protection for gig workers without undermining their flexibility. In Singapore, recent policy recommendations by the Advisory Committee on Platform Workers called for enhanced protection for gig workers who are subject to a “significant level” of algorithmic control.  

Globally, the workforce is crying out for greater flexibility at work, including those in standard employment. Instead of creating specific guardrails for gig workers, we should elevate core worker rights for all, which include stronger digital and data rights. This will go a long way in future-proofing workers against more widespread and harmful algorithmic management practices.