ByteDance’s TikTok Deal Leaves Algorithm Control Question in Limbo
TikTok’s powerful content recommendation algorithm is once again drawing scrutiny after its Chinese parent company, ByteDance, agreed to form a joint venture that will transfer operational control of the app’s U.S. business to American and international investors, including Oracle. The move is designed to prevent a potential U.S. ban and reduce geopolitical tensions, but questions remain over who truly controls TikTok’s most valuable asset — its algorithm.
Who Controls TikTok’s U.S. Operations?
The newly signed agreements mark a significant step towards satisfying Washington’s data security concerns. However, analysts say it is still unclear whether ByteDance has relinquished ownership or control of the recommendation algorithm itself.
Rush Doshi, a former National Security Council official under U.S. President Joe Biden, noted on X that the deal’s details remain opaque, with Oracle appearing to have a monitoring rather than an ownership role.
Sources told Reuters that under the new structure, ByteDance will retain ownership of TikTok’s U.S. business operations but will hand control of data, content, and the algorithm to the joint venture. This venture will manage the back-end operations, overseeing U.S. user data and algorithmic functions, while ByteDance will continue to own the revenue-generating arms such as advertising and e-commerce.
The ByteDance-controlled entity will share part of its earnings with the joint venture in return for technology and data services. Beijing has yet to comment on the deal, but its 2020 export regulations — which require state approval for transferring algorithms or source codes abroad — could complicate any attempt to finalise the arrangement.
Why the Algorithm Matters
TikTok’s success lies in how its algorithm interacts with short-form videos. Unlike platforms such as Meta’s Facebook and Instagram, which depend heavily on users’ social networks, TikTok’s system identifies and refines “interest signals” — patterns of engagement that reveal what each viewer finds compelling.
This interest-based model enables the app to adjust in real time, detecting even subtle shifts in user behaviour, such as what type of content someone prefers at different times of day. Combined with its mobile-first design and early dominance in short videos, these factors helped TikTok outpace competitors like Instagram Reels and YouTube Shorts, which entered the market later.
What Research Shows
Academic studies have shed light on how TikTok’s algorithm works beneath the surface. Research conducted by teams in the U.S. and Germany found that 30% to 50% of TikTok’s recommended videos fall outside a user’s known interests. This deliberate exploration, according to the paper “TikTok and the Art of Personalization,” allows the system to test and refine user profiles, improving retention by surprising viewers with fresh content.
By continuously mixing familiar and exploratory videos, TikTok keeps users engaged longer than most rivals — a strategy that has proven central to its explosive global growth.
Still, whether control of that algorithm will remain in Beijing or shift to the new U.S.-based venture remains the defining question of the deal’s future.
with inputs from Reuters
