Crime and Punishment of a Search Engine

0512-00178-001b1

Does a search engine have the ability to kill a person? After being bombarded by the news about the death of a 21-year-old college student called Wei Zexi, many Chinese have come to the conclusion that it does.

Like numerous online debates about scandals of late, the incident devolves into an exercise of guilt rationing on a massive scale. With the absence of an impartial arbitrator, public opinion takes up the role of fact-finding and responsibility allocation, with implications hard to pin down at this very moment. The story puts Baidu, China’s largest search engine company, at the epicenter of the controversy, bearing the brunt of online criticism, which is guided as much by a complex set of moral convictions as by a vision of technology’s role in the society.

According to Wei’s own account posted on zhihu.com two months before his death (as an answer to the question “what is the greatest evil in human nature?”) , he was diagnosed of a rare tumor, synovial sarcoma. Major hospitals he visited all threw up their hands and told him no effective therapy was available. Desperate, he resorted to Baidu, and initial searches quickly rendered amazing results: a bio-research center based in one of Beijing’s well-regarded hospitals (affiliated with the People’s Armed Police) claimed that they had an advanced therapy (DC-CIK) that could help. The doctor there told him it was “Stanford technology” and promised to extend Wei’s life by “another 20 years at least”. The family invested almost its entire fortune into this last ditch effort, only to find that cancer quickly spread to his lung. Later, well-intentioned individuals on the internet helped Wei find out that DC-CIK was a shelved technology in most parts of the developed world due to limited effect in clinical application. Yet precious time and money was wasted. Wei, the only son of a Xi’an family, died on Apr 12th.

The personal tragedy of Wei Zexi puts a key business component of Baidu under a scorching national spotlight. It is called P4P (pay for performance), whereby customers bid for premium advertisement placement alongside “natural” search results of selected keywords. Although other factors such as quality of content also affect positioning of promoted links, bidding price carries significant weight in the formula, giving high-paying customers good chance of occupying prime locations on Baidu’s search page. The search engine does put a “promotion” mark under sponsored search results, but in a way that is probably not as visually distinguishable as critics and regulators want. The subtlety of the mark can get lost on eyes less experienced with Internet surfing, or those who are eager to find something.

With this background, it may be understandable that the first wave of criticism came for Baidu, even though in both Wei’s original account and the initial investigative piece that directed public attention to the case, the blame fell squarely on the Internet company and on the bio-research center, as well as the invisible yet mightily present state that loomed over the two.

In an era when books like Nudge populate bookstore shelves and people believe in step-counting mobile phone apps to keep themselves fit, the idea that search engine results determine the fate of individual users is only the natural offspring of a faith in the efficacy of technological interventions. It is further enhanced by the towering image of the do-no-evil Google, whose upholding of “enlightened” technology becomes a shining exemplar that shapes the Chinese public’s view of Baidu.

So the conversation swings back and forth between Baidu and Google. Some goes so far as to suggest that Baidu is the “fundamental culprit in dragging down the informational infrastructure of the Chinese society”, by abusing its virtual monopoly in the search market to set up roadblocks on the information highway, profiting from a slowed traffic and a misguided crowd. Google’s Adsense, its core advertising instrument, is upheld as being non-intrusive and responsible. Tales of Google’s efforts to ensure the quality of medical-related search results attracts the attention (and imagination) of Chinese netizens. Very specific ideas proposed by prominent opinion leaders, such as listing ads in a separate column on the screen, are clearly influenced by widely held perceptions of Google’s practices. But it is worth stressing that nowadays Google also puts some ads in the same column as natural search results (with clear marking as “Ads”). More sophisticated industry observers have also pointed out that the growth of Google’s business in China, back when it was still allowed to operate inside the country, was also partially driven by traffic generated by the same kind of search result tricks that Baidu deploys.

Pressed for a response, the Internet firm released a statement through one of its Weibo accounts on Apr 28: it had double-checked the paperwork submitted by the hospital and found it completely legit. This may be true, if your scrutiny stays at hospital level. Move one level below, to the department level, disturbing signs start to emerge. When investigative journalists dug deeper into the bio-research center, they came up with a shadowy web of private entities that had basically “taken over” lucrative departments in military-affiliated Chinese hospitals and ran them like joint-ventures. The public would learn of a so-called “Putian clan”, a group of businessmen who shared the same origin in Putian, a town in southern China’s Fujian province. Lurid, unverifiable stories about the ascent of this group of medical entrepreneurs spread widely on the Internet. As the story goes, they got hold of their first bucket of money in the early days of China’s economic reform. In those years, guerilla clinics prospered in street-side budget hotels, ripping off patients of venereal and skin diseases who were too ashamed to go to proper hospitals. With initial capital in hand, those “bare-foot doctors” began to eye more systematic, legitimate ways of money making. Cash-hungry public hospitals became their natural partners and a new model of “contracted departments” spread like wild fire. In order to bring in more patients, the Putian businessmen took up online marketing, taking advantage of the stellar reputation of hospitals that were hosting them. Baidu’s emergence as a dominant search engine and its offering of P4P handed them a perfect platform to reach out to an anxious, sometimes desperate, clientele. In the process, many patients like Wei Zexi fell victim to sub-par treatments.

The entering of the Putian businessmen into the scene makes the ethical water of the Wei Zexi case much muddier. How much blame should a search engine share if much larger malign interests are motivated to take advantage of its playbook and win access to premium ad slots? As Baidu has always claimed, it only collects and sorts information, not generating it. It acts like a mirror: the reflection is only as good as the Chinese society can be.

An event in 2015 seems to indicate that relationship between Baidu and the Putian clan is less than amiable. At that time, an industry group representing Putian medical interests called for a boycott of Baidu P4P services, claiming that the latter had used its dominance in the market to rip off Putian-controlled hospitals by unilaterally raising prices for promotion. Baidu search results have become a major channel through which such hospitals bring in patients. According to the industry association, P4P expenditures occupy an increasing chunk of those hospitals’ profits, in some cases as high as 60-80%. One the other side of the equation, medical advertisement makes up to 25% of the search engine’s ad revenue, making the relationship of the two parties a love-hate symbiosis. Baidu’s account for the unpleasant stand-off was completely different. It claimed that the industry group was threatening boycott because its crackdown on deceptive medical advertisement was hurting Putian interests. “The threat will not soften our resolute to keep false medical information out of our search results.”

With more information surfacing about the Putian clan, a push back against Baidu-bashing quickly collects momentum. People begin to question the society’s proclivity to blame the safer, easier and more exposed. For them, the focused attack on a publicly listed Internet company is a sign of the collective laziness of Chinese Internet. “The Putian businessmen are happily off the hook now,” as some would proclaim. The despicable deal between public hospitals and the nouveau riche, and negligence of their supervisors, can easily escape public scrutiny under the cover of an outcry directed entirely at Baidu. As a veteran Caijing journalist puts it, what the public should really chase is the regulators who turned a blind eye to the rampant, irresponsible monetization of public hospital reputations. Not only is Baidu a minor consideration in this whole scheme that condemns Chinese patients, so is the Putian clan, whose fortune is determined by the whims of powerful regulators. He predicts that a campaign-style crackdown on private interests in the medical sector would ensue to placate the public, without touching the fundamentals that have allowed the situation to spread in the first place. “It is a way for power to routinely discipline private interest groups, preventing them from growing too big, while reminding them to be more active in paying their rent.”

More methodical minds try to lead people out of this ethical swamp by actually ranking the relative moral responsibility of the parties involved: the biggest share of blame goes to the military-affiliated hospital that knowingly sold its reputation and standing for profit, while being in the best position to judge the medical merit of the technology that its “contracted” bio-research center is promoting. Second comes the center and regulators. Baidu ranks at the bottom of this ethical ladder, for “it is also in relative disadvantage when it comes to medical expertise”. Its only problem is choosing to pursue profits in this category in spite of its own blind spot.

But there are people who resist this way of assigning responsibility. They see it as a distraction or even an intentional tactic to deflect pressure from Baidu, at a point when intensive public questioning is just about to make a dent on one of China’s largest internet firms. The sentiment roots in a deep frustration over a string of Baidu-related controversies (including the one in January this year where it attempted to sell off management authority of an online patient support group to commercial interests), which the Internet giant have all weathered with impunity. “We don’t have the ability to change the root cause of the problem, but at least we can change Baidu with a concentrated effort.” This line of argument contains, at once, a deep sense of powerlessness and a great faith in public opinion: criticizing the power behind the whole corrupt situation won’t bring you much change. It’s a dissipation of precious energy. But the search engine will ultimately bow to such public pressure.

The powerlessness manifests itself in a different reading of Wei’s death. What killed him seems to be a carefully weaved web of sub-lethal elements: acting individually, no element, whether it’s the search engine or the hospital, is potent enough to bring death to a person. Yet collectively, their grip turns out inescapable for an ordinary Chinese like Wei. In the end, each individual element can deny accountability for the collective consequence.

On May 10, China’s Internet authority handed down its verdict on Baidu: it has to change its algorithm for search result presentation, give more weight to credibility, less to bidding prices. Plus, no more than 30% of a page should be given to promotional results. It is a rare occasion where the country’s web regulator publicly dictates change, albeit a noble one this time, to an Internet company’s core business, its algorithm. Earlier, commentators already made a careful-what-you-wish type warning about a more empowered Internet police coming out of this case. But for most part of the cyberspace, vindication is the predominant mood.

The complexity of Baidu’s response to the whole saga is best captured by an article published on the company’s intranet days before the final result. While pledging to collaborate with regulators, it also questioned why the bio-research center could obtain all the certificates and official documents. “As a great enterprise, we sometimes have to shoulder responsibilities that once belong to the state and the medical industry, because with more power comes more responsibility.”

Advertisements

People’s Uber

Uber

It is probably more than a simple play of words that Uber China names its low-end line of service “People’s Uber” (人民优步). The egalitarian connotation sets it in contrast with its Chinese competitors, who invariably give their equivalent services a sense of privilege, calling them “exclusive car services” (专车).

Egalitarian or not, all these smartphone-based, internet-powered car-hailing services are subversive forces that are now sweeping through Chinese cities, upending the old, inefficient order of taxis and their government-backed companies.

In recent months, the struggle between the old order and the new forces has escalated into a national feud with the outbreak of taxi driver strikes in many Chinese cities, violent clashes between taxi drivers and Uber drivers, and ground-breaking lawsuits that try to define the legality of such services.

It is a situation that proves thorny for the government. As recent cases show, Uber and other copycat services are so wildly popular with an urban, tech-savvy class profoundly dissatisfied with the taxi services that any moves against the Ubers induce vocal discontents online, at a time when the top leadership is trying hard to woo the support of this group. But on the other hand, the interests that Uber threatens are also highly organized and entrenched. Their response to the advance of the Ubers can easily turn massive and nasty.

Two cases last week vividly illustrate the above dilemma. On May 21 the Beijing transportation administration carried out a major crackdown on what they considered “black taxis”, which involved private cars running Uber-like services. When pictures of the crackdown emerged on the internet, showing enforcement officers using hammers to forcefully break into cars and drag out drivers, they were greeted with an outburst of indignation and ridicule. Liberal commentators questioned the legality of such actions; Public figures, including well-known stand-up comedian Joe Huang and movie star Yao Chen, used their Weibo accounts to express discontent and disbelief. Tough-in-the-cheek Weibo personalities readily made fun of the authority by teasing them about “why not using the anvil as well.”

On the same day in Tianjin, a mega-city two hundred kilometers away from Beijing, taxis drivers took matters into their own hands. One of them solicited an “exclusive car service” as a customer, and then brought the car to a pre-arranged destination, where a group of taxi drivers waited to “give the driver a lesson”. Apparently, the Uber driver was not easily intimidated. He sent out a WeChat alert to his community of similar drivers, and in no time, hundreds of them came to his rescue. A major stand off and violent clash quickly ensued, which in the end required police to help disperse.

Caught between the wild popularity of internet car-hailing services among urban middle class consumers and the intransigent taxi interests, the central government’s position so far has been stubbornly ambiguous, if not outright self-contradictory. In Nov 2014, the spokesperson for China’s Ministry of Transportation publicly declared that internet-based car services “should not be killed with one strike of the rod.” It was widely read as a half endorsement of the Ubers until, four months later, the Minister himself hardened the Ministry’s stance by famously saying that “private vehicles should NEVER be allowed into the commercial services”, citing safety concerns as the main consideration.

But Uber and its sister services continued their aggressive expansion unabated. Part of the reason might be the existence of a mixture of seemingly authoritative signals which are subject to all kinds of favorable interpretations. Only recently, China’s Premier Li Keqiang has been advocating the concept of “Internet+”, a vision to upgrade China’s economy to a more information and data powered platform. And early this year, while numerous local governments were ramping up their campaigns against the Ubers, China’s central news agency Xinhua and the party’s no.1 propaganda outlet People’s Daily were publicly supporting such services with a series of widely circulated opinion pieces. For a nation so accustomed to reading the “tea leaves” of the coded signals from the top, such an environment is fertile for over-enthusiastic (mis)judgments.

Kenneth Lieberthal, an influential China expert in the US, once observed that it is difficult for outsiders “to determine with confidence which outcomes reflect strategic decisions by China’s national leaders and which instead reflect inherent dynamics of the political system that are beyond the control (and some-times against the wishes) of those leaders.” The Uber situation is a typical illustration of this challenge as at times you can’t tell whether matters are manifestations of the political will from the top, or mere reactions prompted by local power dynamics. Take the above mentioned Tianjin for example, on the one hand the municipal government announced all internet-based car-hailing services “illegal” a few months ago, on the other hand it has thenceforth only penalized eight “illegal” vehicles. It is likely that such actions were mere tokens to placate the protesting taxi interests, which were particularly vocal in Tianjin. There the ownership of taxis belongs to individual owners rather than large taxi companies, the former being much more financially vulnerable. Such a unique local set-up produces interactions that project uncertainty onto the national level where every development is closely watched by many.

The raw subversive power of Uber is creating problems for regulators all over the world. While the Chinese authorities are not the only ones scratching their heads, they do face a peculiar challenge of lacking institutionalized means to negotiate a legitimate compromise with the numerous interests involved. When an official from China’s central planning agency (NDRC) got interviewed for the Uber question, he made a typically dialectic instruction to the local governments: seriously take into account the need of the public, while at the same time handle the historically accumulated problems (i.e. taxis) with grace.

Not very instructive.