The latest goal of China’s tech regulatory lightning: algorithms
In this illustration photo taken on July 12, 2017, computer code can be seen on a screen over a Chinese flag.
Thomas White | Reuters
BEIJING – Chinese authorities plan to restrict companies’ use of algorithms in selling products to consumers, which analysts say is likely to be counter to business interests and set a precedent for other countries.
China’s largest tech companies, from e-commerce giant Alibaba to TikTok owner ByteDance, have built their multi-billion dollar businesses on algorithms that deliver content that a customer is more likely to spend money or time on based on previous views.
The regulatory authority for cybersecurity, which is becoming more and more powerful, published far-reaching draft rules last Friday to regulate the use of these so-called recommendation algorithms. The proposal can be commented on until September 26th, no implementation date has yet been set.
The groundbreaking rules could lead to a clash between China’s tech giants – which have come under increasing regulation over the past 10 months – and Beijing, which has tried to curb its power.
And China’s algorithm rules are being closely watched by other countries and technology companies around the world to learn how they could affect business models and innovation, analysts said.
“Companies will have a lot to say about this because it has the potential to restructure business models,” Kendra Schaefer, partner at Beijing-based consulting firm Trivium China, told CNBC.
The rules have also raised questions about how enforcement will be done and how intrusive regulators need to be to actually get companies to comply with these rules.
What the design says
Here are some of the key points in the draft rule:
- Corporations must not set up algorithms that induce users to become addicted or to spend large sums of money.
- Service providers must clearly inform users about the algorithmic recommendation services they provide.
- Users must have a way to turn off algorithmic recommendation services. Users should also have the option to select, revise, or delete user tags that will be used for the recommendation algorithm.
- If algorithms are used to market goods or provide services to consumers, the company behind them may not use the algorithm to make “unreasonable” differentiations in relation to prices or trading conditions.
- Any violation of the rules could penalize companies with fines ranging from 5,000 to 30,000 yuan ($ 773 to $ 4,637).
These proposed rules come as the Chinese government tightened its regulation on domestic tech giants last year, mostly in the name of fighting monopoly practices and increasing privacy.
A new data protection law came into force on Wednesday. On November 1st, a law on the protection of personal data comes into force.
This is what enforcement could look like
Recommendation algorithms consist of code that is fed specific information about users to enable tailored provisioning Results. When you’re on an ecommerce website, some of the items you see on the home page are likely there because of your browsing or shopping habits.
However, the code of the algorithm will not be published and could make enforcement difficult. At the very least, it could require regulators to review the code of the companies behind the algorithms.
“You can’t do algorithmic regulation without looking at the code,” said Trivium China’s Schaefer.
Authorities are expected to conduct algorithm “security assessments” and inspections of referral services according to the draft rules. Companies must cooperate and provide the necessary technical or data support.
That would give tremendous power to regulators in China.
But it also poses some challenges.
“First of all, you need the technical capacities.
This intrusiveness could lead to a clash between China’s tech giants and regulators.
“I’m sure companies have privacy rights issues … that [the code] is protected information, “added Schäfer.
Neither Chinese tech company CNBC contacted immediately commented on the draft rules, with two indicating that it is too early in the process to evaluate them. The cybersecurity regulator did not immediately respond to a CNBC request for comment on the scope of implementation or the impact on innovation.
Changes in the business model?
Many of the Chinese tech giants don’t make money directly from their algorithms. Instead, they are used to attract consumers to products. For example, you can watch videos in an app and get recommended similar content. A business would monetize this through advertising, or even trick you into buying things.
The latest rules could force companies to change their business models, but to what extent is unclear.
“The jury has not yet decided on the impact on operations and profits,” said Ziyang Fan, director of digital commerce at the World Economic Forum.
“It depends on a number of factors, such as the level of enforcement and market reaction – how many users would choose to ‘shut down’. [the] Recommendation algorithm if this leads to a sub-optimal user experience, e.g.
“When we see a significant drop in indicators like DAUs [daily active users] and retention rates, then the impact on profits could also be significant, “he said, noting that social media companies may see the impact more strongly while online shopping and ride-hailing are” likely less “.
Where the rest of the world is
As the interface between technology and daily life grows, countries and regions around the world are increasingly looking for ways to regulate technology and the companies that sell it.
So far, this has led to different approaches. In the area of algorithms, China is specifically focusing on the recommending function of the technology, while the US and the European Union are discussing broader laws surrounding artificial intelligence.
At the beginning of this year, the European Union tabled a bill called Artificial Intelligence, which aims to facilitate the “development of a single market for legitimate, safe and trustworthy AI applications” and to promote innovation in this area.
The law contains “specific requirements aimed at minimizing the risk of algorithmic discrimination”.
But there are a number of differences from China’s algorithm rules.
Fan of the WEF said the EU is taking a “risk-based approach” while China’s rules “do not distinguish between levels of risk and apply to all use of algorithm recommendation technologies.” This can cover a wide range of industries, from grocery delivery to education.
And China’s rules “target algorithms directly at the user and product level,” such as the ability for users to turn off the algorithm as stated in the proposed rules, Fan added.
Read more about China from CNBC Pro
Once China’s algorithms law goes into effect, it will be closely watched around the world as authorities try to figure out how to regulate the technology in the future.
“That will be a global example,” said Schäfer. “Tech companies overseas will see how Chinese tech companies benefit or not in the face of these limitations on algorithms. If they change their business models, if they can succeed in spite of the regulation of algorithmic processes, there is very little excuse for … foreign governments . ” not to do the same. “
“If they fail and aren’t that profitable and shareholders are disappointed, that’s bad too,” she said. “This supports the argument that algorithmic regulation cannot be implemented without detrimental effects on innovation.”