Thread by @_KarenHao

Governments around the world are grappling with how to stamp out misinfo & toxicity on the internet. China is first to advance its solution: an unprecedented plan to control platforms' underlying algorithms.

But there's a problem. It may be impossible. wsj.com/articles/china…

Earlier this month, China's top internet regulator announced that two dozen of the country's most influential internet companies had submitted a total of 30 of their core recommendation algorithms.

These included Bytedance's personalization algorithm for Douyin (the Chinese version of TikTok), Baidu's search-ranking algorithm, and Alibaba's Tmall product-recommendation algorithm. The full list is translated here. chinalawtranslate.com/en/algorith-li…

The move is meant to be the first step in the regulator's process to start enforcing a regulation that came into effect in March, which noted ambitions to stamp out the amplification of harmful information and user privacy infringements. digichina.stanford.edu/work/translati…

The idea is enticing. If algorithms are wreaking havoc and companies are using them to evade responsibility, why not increase transparency into what they're doing and finally put in some guardrails.

FB whistleblower Frances Haugen effectively argued for a version of this last fall when she pushed US & EU regulators to place some kinds of limits on social-media algorithms. Other orgs have pushed for greater algorithmic transparency. China is taking both ideas to the max.

It's left some tech experts feeling jealous that China is actually moving to close what they see as a gaping regulatory hole that still exists in the US. As one put it to me, "algorithms are the perfect mechanism to bypass regulation."

But US & EU regulators also have good reason. Regulating algorithms so directly could set a precedent for greater state control over the flow of information. Indeed China's law includes a provision that algorithms be used to promote content that treats the government favorably.

That leaves China's effort as the sole attempt globally that is testing this proposition: Can you tame the internet by taming the automated systems that govern it?

The answer might be no.

The trouble is that today's internet algorithms are largely based on machine learning. And apps like Facebook and TikTok use not just one but hundreds of machine-learning algorithms.

Even if you had highly detailed documentation or the code to these systems, it wouldn't give you full insight into how they work. What matters is the data that's running through them—and on social media, that data is the ever-changing stream of user-generated posts and behaviors.

That means targeted changes like promoting more propaganda are feasible. “But it is actually impossible to control what a recommendation engine does overall,” says Cathy O'Neil, an algorithmic auditor who works with U.S. government agencies to scrutinize company algorithms.

That said, China's efforts are important to watch. As @gwbstr told me: "They are doing things that no one else has tried yet, and the rest of the world can learn from what works and doesn’t work."