October 6, 2025

Trump’s agreement will fundamentally change the Tiktok algorithm for you in the future, social media

0
GettyImages-2194585249-e1758558200208.jpg


President Donald Trump announced on September 19, 2025, a preliminary agreement for the sale of majority participation in Tiktok of the Chinese technology giant bytedance in a group of American investors after Trump’s negotiation with the Chinese leader Xi Jinping.

The agreement would create a new American version of the application, bringing it into compliance with a law signed by former president Joe Biden on April 23, 2024, and confirmed by the Supreme Court on January 17, 2025. The details of the agreement remain to be developed, and what this means unresolved is the fate of the basic algorithm of the application of the video.

The Chinese government has indicated that it will not allow Bytedance to sell the algorithm, because it is classified as controlled technological export, according to Chinese law. Meanwhile, the leaders of the American technology industry and some legislators say that compliance with the law requires algorithm to be under American control. The agreement as proposed includes the license of the algorithm so that it remains a Chinese intellectual property while the American version of the application continues to use the technology.

Tiktok page algorithm is largely considered the most important part of the application. As an analyst said: “The purchase of Tiktok without the algorithm would be like buying a Ferrari without the engine.”

The value of the algorithm lies in its strange capacity to anticipate user content preferences. Many users say that it knows them better than they know themselves – a feeling that has evolved into a curious mixture of spiritual beliefs and conspiracy theorization, like my colleagues and I have documented. Other researchers have also noted that users feel more intimately seen and known to Tiktok algorithm than those who feed other popular platforms.

I have studied social media algorithms for almost a decade, exploring how our relationships with them have evolved as they become more and more linked to daily life. As a social media and devotee of Tiktok, I want to shed light on the operation of the algorithm and how the application could change in the wake of its sale.

How the Tiktok algorithm works

In some respects, the Tiktok algorithm does not differ considerably from other social media algorithms. Basically, algorithms are only a series of steps used to achieve a specific objective. They carry out mathematical calculations to optimize the release at the service of this objective.

There are two layers in the Tiktok algorithm. First, there is the abstract layer that defines the results that developers wish to achieve. An internal document shared with the New York Times said that Tiktok algorithm optimizes for four objectives: “user value”, “long-term user value”, “creator value” and “platform value”.

But how do you transform these goals into mathematics? What does an abstract concept like “user value” mean? It is not practical to ask users if they appreciate their experience whenever they visit the site. Instead, Tiktok relies on proxy signals that reflect abstract results in quantifiable measures – in particular, I like, comments, shares, follows, the time spent on a given video and other user behavior data. These signals are then part of an equation to predict two key concrete results: “retention” or the probability that a user returns to the site and the “time spent” on the application.

The Tiktok algorithm for your page is based on automatic learning to predict retention and time spent. Automatic learning is a calculation process in which an algorithm learns models in a set of data, with little or no human guidance, to produce the best equation to predict a result. Thanks to learning models, the algorithm determines the amount of individual data signals that this counts to find a precise prediction.

A Wall Street Journal survey has revealed that the time that users spend watching each video plays an important role in the way algorithm chooses videos it suggests to users. Using the equation he has generated to predict retention and time spent, the algorithm assigns a score with each video and classifies the possible videos that could be shown to the user by this score. The higher the score for an individual user, the more the video will probably appear in its flow.

Of course, content characteristics and other users also shed light on the recommendations, and there are other sub-processes folded in the equation. This step is the place where algorithmic moderation generally comes into play. If a video looks like an engagement bait or an excessive gore, for example, the content score will be penalized. Here are the basics of the functioning of the Tiktok algorithm.

Which is likely to change for American users

The sale has not been finalized and what happens to the algorithm is not resolved. However, it is quite certain that Tiktok will change. I see two key reasons for change.

First, the proposed American application population will change the composition of the underlying data set informing the algorithmic recommendations on a continuous basis. As content types and users reflect cultural preferences, American values ​​and behavior, algorithm can be slightly different because it “learns” new models.

In addition, not all users will choose to join the new application, especially if it is considered under the control of Trump’s allies. The current agreement would grant an 80% share of American investors, 50% of which to new investors Oracle, Silver Lake and Andreessen Horowitz. These investors have links with Trump, and an apparent provision of the Agreement allows the US government to select a member of the board of directors. This can lead to a user population – and data – reflecting a narrower area of ​​interests and ideologies.

Second, it is possible that the owners of majority sharing of the new application decide to adjust the algorithm, in particular with regard to the moderation of the content. New owners may wish to modify Tiktok’s community directives according to their point of view of an acceptable and unacceptable discourse.

For example, Tiktok’s current community directives prohibit disinformation and work with independent fact verifiers to assess the accuracy of the content. While Meta followed a similar approach to Instagram and Facebook, in January 2025, he announced that this would end Meta’s relations with independent fact checks and loosen the content restrictions. YouTube also softened its content moderation this year.

The main thing is that algorithms are very sensitive to the context. They reflect the interest, values ​​and visions of the world of people who build them, the preferences and behavior of people whose data inform their models and the legal and economic contexts in which they operate.

This means that if it is difficult to predict exactly what an American tiktok will only look like, it is sure to assume that it will not be a perfect mirror image of the current application.

Kelley Cotter, assistant professor of information and technology sciences, Penn State

This article is republished from the conversation under a Creative Commons license. Read the original article.The conversation

Global Forum fortune returns on October 26 to 27, 2025 in Riyadh. CEOs and world leaders will meet for a dynamic event only invitation that shapes the future of business. Request an invitation.




https://fortune.com/img-assets/wp-content/uploads/2025/09/GettyImages-2194585249-e1758558200208.jpg?resize=1200,600

About The Author

Leave a Reply

Your email address will not be published. Required fields are marked *