In January, a worker at, a Chinese food delivery platform, doused himself in gasoline and set himself on fire in the eastern city of Taizhou to protest withheld earnings. The company had refused to give him his pay, around $700 (5,000 yuan), after he attempted to switch to another platform. Covered in ash and third-degree burns, the man was filmed yelling, “I want my blood-and-sweat money,” as bystanders tried to extinguish the flames. Since then, there have been more protests against China’s food delivery apps: last month, as Rest of World reported, dozens of drivers at’s rival Meituan took to the streets of Shenzhen to protest a new policy that threatened to reduce their wages.

Since the outbreak of the pandemic, the plight of delivery personnel has become a question of global significance. Across the world, the food delivery industry has exploded; in China, Meituan alone recruited 458,000 new drivers between January and March 2020. This unprecedented expansion has been a boon to companies but has also made life even harder for delivery workers. In China, this has triggered a widespread backlash against the country’s app-based delivery industry. A viral investigation published last fall in Renwu magazine exposed shocking details about its inner workings: how algorithms set unrealistic expectations, the ease with which drivers can be fired, and the frequency of casualties and gory injuries. In response to public fury on social media, Meituan and announced new policies — they would give drivers an extra five to 10 minutes to make their deliveries. But they failed to address the deep-seated, structural problems that enable this kind of exploitation: namely, toxic work cultures that maximize competition and the use of increasingly sophisticated algorithms that track, monitor, and govern the lives of employees. 

In China, only 1% of delivery workers make $1,532 a month, while more than half make less than $766.

The low-level workers in China’s $97 billion food delivery sector are especially vulnerable to brutal working conditions and exploitative management practices for several reasons. First, the entire industry is locked in competition between two platforms. Meituan, backed by tech giant Tencent, controls 67% of the market share;, owned by rival Alibaba, holds 31%. Having subsumed all other competitors (including Baidu’s delivery service), the two companies now fight over customers by offering steep discounts and cutting costs to the bone. This sometimes veers into open antagonism: drivers are taught to chant the slogan “Kill Meituan, is fighting with you.” In its early years, Meituan became known for cultivating a strategy of “gladiatorial entrepreneurialism,” which consisted of poaching employees and launching smear campaigns, and deployed “tactics that would make Travis Kalanick blush,” writes Kai-Fu Lee in his book AI Superpowers

Cutthroat competition is not only baked into both companies’ cultures but is also embedded in the algorithms that underlie each platform. Meituan’s “Super Brain” and’s “Ark” systems are driven by vast supplies of data collected on individual users, drivers, and their deliveries. “The more data collected,” explained Jeffrey Ding, an AI researcher at the Oxford Future of Humanity Institute, “the more efficient and precise the algorithms [become].” Over time, these systems get better at allocating tasks, assigning routes, and optimizing deliveries. In turn, this encourages customers to order more takeout, prompts workers to move just a little faster, and provides platforms with even more information. Then, the cycle accelerates. In 2016, maximum delivery times were capped at 1 hour; by 2018, this was whittled down to 38 minutes. 

To further expedite delivery speeds, platforms have also developed a gamified evaluation process that uses a point system to reward “effective” workers and punish “inefficient” ones. Meituan drivers are also ranked from “ordinary” to “king” levels.” Drivers who don’t meet quotas, make their deliveries on time, or receive bad reviews are fined, docked points, and often forced to reimburse the company. To keep up with ever-shrinking delivery windows, drivers routinely speed, run red lights, and go against traffic. This, unsurprisingly, has come at a human cost. In the first half of 2017, one delivery driver died every two and a half days in Shanghai; in 2018, one delivery worker was injured or killed on average every day in Chengdu. 

In contrast to the United States, where drivers are won over with promises of “freedom” and “flexibility,” Chinese workers are sold on “competition” and “upward mobility.” “In China, delivery work is not advertised as a side gig you do for a couple hours,” explained Ding, “but as a way to make a lot of money and make it quickly.” In other words, hustle hard enough, and maybe you’ll rise to the level of “King.” But slack off, and you’ll be leveled down and tossed back into the heap. The dream of earning $1,532.00 (10,000 yuan) a month is a powerful one, and it is what companies use to draw on a seemingly endless supply of potential workers. That’s what attracted A’Fei, who left KFC to join Meituan as a delivery driver. A’Fei told Renwu magazine that, at the beginning, he did hit this target — while working 9-hour days and accepting long-distance jobs that nobody else wanted. But eventually, he was unable to keep up this pace, and his income dropped, sometimes bottoming out at less than $1,072 (7,000 yuan) a month. In reality, the expectation was unrealistic in the first place: only 1% of delivery workers make $1,532 a month, while more than half make less than $766 (5,000 yuan).

China’s food delivery industry is fueled by a large and cheap workforce of 6 million people, most of whom are young men who moved from the countryside to China’s fast-paced cities, in hopes of building better lives for themselves. Because most delivery workers are hired as either part-time or full-time independent contractors, they are not given insurance or compensated for work-related injuries. They also have no real bargaining power, as independent unions are banned in China. The best workers can hope for is a sympathetic employer. Otherwise, they can stage illegal strikes, organize informal mutual aid networks to support each other, and share information about new bonus policies and cheap bike rentals. 

It is important to note, however, that the factors that enable Chinese food delivery platforms to exploit workers are by no means unique to China. Even in places where unions are still legal and viable, powerful tech corporations are learning how to strip workers of their agency. In the United States, for example, Amazon uses surveillance and artificial intelligence systems to assess and monitor workers. According to The Washington Post, the algorithm that the company uses to help workers pack more efficiently is called, unironically, “The Matrix.” Reporting has also revealed that Amazon invested in technology to track and suppress labor organizing efforts at the company and even posted job listings for intelligence analysts to monitor “labor organizing threats” (these have since been removed). Earlier this month, Amazon employees at a warehouse in Alabama voted against unionizing the facility’s workforce. This demands the question: Was this a fair vote? In the end, around half of eligible warehouse workers abstained from voting. Just as Meituan’s drivers are managed by the “Super Brain,” and’s governed by its “Ark,” did Amazon’s workers, under the company’s watchful eye, truly have full agency over their own decisions? 

The factors that enable Chinese food delivery platforms to exploit workers are by no means unique to China.

Today, those who suffer most under systems of algorithmic labor are the most marginalized: in China, that means rural migrants, manual laborers, and young men and women from the lowest rungs of Chinese society. But this may not be the case for long, as white-collar workers are increasingly subject to the same kinds of systems. Many companies rely on office automation software equipped with “intelligent managers” to take attendance, allocate tasks, and “improve staff efficiency.” Alibaba’s workplace app DingTalk, already ubiquitous in companies across China, is equipped with geo-tracking technology to monitor employee whereabouts. In January of this year, a Hangzhou company was exposed for using sensor-equipped cushions to monitor employees’ heart rates and posture and ensure they weren’t slacking off. “I felt like I was naked,” said one employee who found out she was being monitored by a smart cushion. This trend isn’t likely to abate anytime soon: The Chinese market for digital management tools grew from $1,072,041,166 (7 billion yuan) to more than $1,684,636,118 (11 billion yuan) from 2017 to 2019 and is expected to exceed $3,062,974,761 (20 billion yuan) by next year.

As the kinds of automated systems that govern the lives of delivery workers begin to creep into other sectors, it’s necessary to realize that this trend is structural and enabled by factors that are pervasive and not unique to China: namely, the unchecked power of tech monoliths and a toxic work culture that prioritizes speed and efficiency over human life. Without systemic reform, tech companies, in attempting to rid employees of their worst qualities — their “laziness” and “inefficiency” — will end up stripping them of their best ones. Ironically, these are the very skills that are crucial to the success of any company: creativity, an ability to collaborate, and a capacity for trust and empathy.