Hello and welcome back to TechCrunch’s China roundup, a digest of recent events shaping the Chinese tech landscape and what they mean to people in the rest of the world.
The biggest news of the week again comes from Beijing’s ongoing effort to dampen the influence of the country’s tech giants. Regulators are now going after the exploitative use of algorithm-powered user recommendations. We also saw a few major acquisitions this week. Xiaomi is acquiring an autonomous vehicle startup called Deepmotion, and ByteDance is said to be buying virtual reality hardware startup Pico.
Beijing has unveiled the draft of a sweeping regulation to rein in how tech companies operating in China utilize algorithms, the engine of virtually all lucrative tech businesses today from short videos and news aggregation to ride-hailing, food delivery and e-commerce. My colleague Manish Singh wrote an overview of the policy, and here’s a closer look at the 30-point document proposed by China’s top cyberspace watchdog.
Beijing is clearly wary of how purely machine-recommended content can stray away from values propagated by the Communist Party and even lead to the detriment of national interests. In its mind, algorithms should strictly align with the interest of the nation:
Algorithmic recommendations should uphold mainstream values… and should not be used for endangering national security (Point 6).
Regulators want more transparency on companies’ algorithmic black boxes and are making them accountable for the consequences of their programming codes. For example:
Service providers should be responsible for the security of algorithms, create a system for… the review of published information, algorithmic mechanisms, security oversight… enact and publish relevant rules for algorithmic recommendations (Point 7).
Service providers… should not create algorithmic models that entice users into addiction, high-value consumption, or other behavior that disrupts public orders (Point 8).
The government is also clamping down on discriminative algorithms and putting some autonomy back in the hands of consumers:
Service providers… should not use illegal or harmful information as user interests to recommend content or create sexist or biased user tags (Point 10).
Service providers should inform users of the logic, purpose, and mechanisms of the algorithms in use (Point 14).
Service providers… should allow users to turn off algorithmic features (Point 15).
The regulators don’t want internet giants to influence public thinking or opinions. Though not laid out in the document, censorship control will no doubt remain in the hands of the authorities.
Service providers should not… use algorithms to censor information, make excessive recommendations, manipulate rankings or search results that lead to preferential treatment and unfair competition, influence online opinions, or shun regulatory oversight (Point 13).
Like many other aspects of the tech business, certain algorithms are to obtain approval from the government. Tech firms must also hand over their algorithms to the police in case of investigations.