Big tech companies such as Meta (Facebook), Google, Amazon, Twitter, and others have always tried to keep their algorithms close to their hearts. At the end of the day, these long stretches of code are main contributors to the company’s revenue streams. Whether it is in marketing, advertisement selection, recommendation, search optimization, or other functionalities, the underlying algorithms allow big techs to optimize their content both to increase their user base and generate an increasing flow of money. What happened in Japan between restaurant chain Hanryumura and restaurant review platform Kakaku.com has opened the door to a sequence of potentially bold decisions that could impact big tech companies inside the country and later outside.
The prospect of forcing big tech companies to disclose their algorithms would certainly be a major blow to their operations, and a defining factor in determining how competition would unravel in the future with considerably fewer secrets left under the blanket.
The Trigger to the Decision: Hanryumura VS. Kakaku.com
Korean BBQ chain Hanryumura has witnessed in May 2019 a drop in their review scores by an average of 0.2 points with some stores having a decline that go as much as 0.45 points. For a large restaurant chain with 21 outlets and a scoring scale between 1 and 5 (5 being the best), such a drop is considered significant. The lower rating has affected the chain’s sales as the Tabelog website, operated by kakaku.com Inc., is considered one of the leading restaurant guides in Japan with more than 800 thousand restaurants and a number of reviews exceeding 46 million. Hanryumura has attributed the slump in sales to an implicit algorithm change by the website operator kakaku.com which targets large restaurant chains in order for those to have more expensive contracts with the review website.
The court in Japan acknowledged that the operator of the Tabelog website, kukaku.com violated the anti-monopoly law abusing its position to make unilateral changes to its rating mechanism without communicating these modifications to its member restaurants. Eventually, the defendant had to pay around USD 286 thousand in damages to the restaurant chain, a sum which is however much lower than what the plaintiff requested.
More importantly, the game-changing decision made by the Japanese court was the kakaku.com had indeed to disclose part of this algorithms, especially that related to how rating scores are being calculated.
Disclosing Algorithms May Reveal Trickery
The controversial decision to open up and disclose algorithms details is the worst nightmare big tech companies may have to live with in the future. While kakaku.com is a big company in Japan, the amplitude of the decision on international giants such as Meta, Google, or Amazon would be notably much bigger. Revealing the algorithm is like trying to discover a cave. The companies themselves do not understand fully how the code is delivering the intended results. The situation is similar to the relation between traditional AI algorithms and explainable AI. When interpretability becomes a core component in the design of AI algorithms, the scope of the approach becomes completely different as sources of bias can be avoided. Therefore, companies may have trouble explaining the philosophy behind their algorithms. Digging deep into the code could even emphasize some trickery. This is a common trend in research and development as companies intentionally inject bias in their code to meet their projected targets.
The biggest impact of disclosing algorithm details is however that competitors can build on a successful product to even produce a better one. Instead of having the other platform as a benchmark, disclosing the algorithm makes it an open-source code that can be upgraded.
Disclosing Algorithms May Raise Security Concerns
Making algorithms public has another dark side which relates to security. Divulging details is a feast hackers are eagerly waiting for. The access to the full code allows hackers to look for loopholes and backdoors. Since marketing and ratings depends on private customer data, hackers many even find a way to steal sensitive information which is essential for us consumers to protect.
Opening up on algorithmic details will undoubtedly pave the way to a myriad of cybersecurity attacks and concerns, and consequently notable headaches for the big techs which would need to invest more into their cyber-defense mechanisms.
A Snowball Effect
What happens in Japan would certainly not stay in Japan. The large geographic footprint that big tech companies have could be an unwelcomed source of concern. Legislators around the world would learn from the Japanese experience to have specific requirements regarding the companies’ algorithmic details. Therefore big techs are in for a headache if they agree to disclose algorithms.
The snowball effect is not of geographic dimension only, but rather could escalate to include all websites and platforms that use data for specific targets. Not only big tech companies will be affected, but rather all websites that collect reviews, personal information, and act accordingly.
The Chinese experience with its Splinternet could extend to other regions as well with big tech companies having to adapt their offerings, cybersecurity measures, data-sharing practices, and even developed algorithms as per the law governing online operations.
A small event, yet with much greater ramifications. This is the clear description of the lawsuit between Hanryumura and restaurant review platform Kakaku.com. The bold decisions made by the Japanese court to stand by the plaintiff side and even request part of the online platform’s algorithm to be disclosed is an event by itself. The drive to more transparency will be contagious to big tech companies which will have to have major shake-ups to abide by the emerging requirements in Japan and outside. This will consequently create additional challenges these companies have been trying to avoid.
Inside Telecom provides you with an extensive list of content covering all aspects of the tech industry. Keep an eye on our Technology space to stay informed and up-to-date with our daily articles.