The TikTok app is under growing pressure as the deadline for selling itself in the US is drawing closer. Meanwhile, they are also under pressure from China when it passes a decree restricting the export of artificial intelligence products, an area related to the algorithms within the application.
While seeking to negotiate with the US government to find a more viable solution, TikTok executives have taken a bold step to eliminate concerns about their companies’ connections to the government. China: revealing its “magic” algorithm – which attracts hundreds of millions of users around the world to focus on its cult video application.
TikTok management said that the revelation of the details of their algorithm and how it processes data is intended to shatter the rumors and myths about its company:
“We are a 2 year old company operating with the expectation of a 10 year old company“Michael Beckerman, TikTok vice president for public relations in the US,” said.We didn’t stand a chance to grow up in the golden age of the internet, when tech companies couldn’t do anything wrong. We grew up in an age of tech clash, where there is always a lot of skepticism about platforms, how they censor their content and how their algorithms work.. ”
TikTok management also allows reporters to take a virtual tour of the “Transparency Center” in Los Angeles. This center has areas to demonstrate a demo of TikTok’s algorithm and how the company processes data.
How does TikTok’s algorithm work?
According to the company, TikTok’s algorithm uses machine learning to determine which content will be more likely to interact with a certain user and show them more of that type, by searching for more. videos are like that as well as users with similar interests.
When users open TikTok for the first time, they will be shown 8 popular videos, characteristic of different music trends, topics. This algorithm then continues to show new users eight other videos based on the videos the user interacted with and what the user did.
This algorithm identifies similar videos based on its information, including details like its captions, hashtags, or audio. Video recommendations are also device-based in relation to the user account and its settings, including data such as language preferences, country setting, and device type.
When TikTok collects enough data about the users, the application maps user preferences in relation to similar users and groups them into “clusters”. Also, the app groups videos into clusters based on similar topics, like “basketball” or “rabbit”.
Using machine learning, the algorithm renders incoming videos to users based on how close they are to other user groups and what content they like.
TikTok does this to avoid users getting bored of having to watch multiple videos with the same music continuously or from the same creator.
TikTok acknowledges that their algorithm is so appealing to users’ interests that it can create a “filter bubble”, reinforcing the user’s existing interests rather than showing them more diverse content, broaden their horizons or give them opposing perspectives – which can reinforce conspiracy theories or false information. So the company says it’s working to find ways to break them when needed.
For user data
The company says it’s trying to classify and prevent crashes on its platform before they happen. To do that, they research and discover problem patterns before they spread.
TikTok chief security officer, Roland Cloutier, said it plans to hire more than 100 data, security and privacy experts by the end of this year in the US.
He also said that the company will build a monitoring, response and investigation center in Washington to proactively detect and respond to serious incidents in real time.
Refer to Axios