ByteDance, owner of TikTok, has confirmed it terminated an intern in August for “maliciously interfering” with an AI model’s training while disputing widespread social media claims about the extent of the damage.

According to a BBC report, the company stated that the intern, who worked on the advertising technology team, had no access to or experience with ByteDance’s AI Lab. The firm explicitly denied reports that the incident disrupted a GPU-based AI training system and caused more than $10 million in damages.

“Their social media profile and some media reports contain inaccuracies,” ByteDance said in an official statement, pointing out that its commercial operations, including large language AI models, remained unaffected by the intern’s actions.

Following the incident, ByteDance notified both the intern’s university and relevant industry bodies. The company maintains its position as a leading AI developer, mainly through its Doubao chatbot, which remains China’s most popular AI chatbot platform.

ByteDance continues to expand its AI capabilities across multiple applications, including Jimeng, a text-to-video tool. It also maintains its position in social media through TikTok and its Chinese counterpart, Douyin.

The clarification comes as ByteDance, like many global tech firms, increases its investment in AI technologies across its platform ecosystem.

The company reportedly plans to use Huawei chips to develop a new AI model as China intensifies efforts to reduce dependence on U.S. technology amid tightening export controls. ByteDance allegedly ordered over 100,000 Huawei chips but has received fewer than 30,000 as of July.

Reports emerged that ByteDance’s web crawler, Bytespider, is scraping web data approximately 25 times faster than GPTbot, the crawler used by OpenAI for its ChatGPT platform. However, the bot reportedly doesn’t respect robots.txt; a code website publishers use to signal that their data should not be collected.

This month, TikTok announced it was laying off approximately 500 employees in Malaysia as part of a global shift towards increased use of AI in content moderation.