logo80lv
Articlesclick_arrow
Research
Talentsclick_arrow
Events
Workshops
Aboutclick_arrow
profile_loginLogIn

An Intern Sabotaged TikTok's AI Models, Allegedly Costing ByteDance Millions

ByteDance confirmed the sabotage but denied losing tens of millions of dollars because of it.

In yet another sign that the conflict between humans and artificial intelligence is already unfolding globally, a rumor spread across the Chinese part of the internet over the past weekend, alleging that an intern at ByteDance, the tech giant behind TikTok and its Chinese counterpart Douyin, sabotaged the company's AI models, causing tens of millions of dollars in damages.

Jonathan Raa/Getty

According to the rumors, the intern's sabotage involved planting a malicious code to halt and/or slow down TikTok's AI model training, purportedly powered by over 8,000 NVIDIA H100 GPUs. Because of this interference, ByteDance reportedly lost over $10 million, factoring in the loss of computing power, wasted time, and wages of the AI research team.

In just a couple of days, the gossip got so loud it exited the Chinese corner of the World Wide Web and started making headlines all over the globe, prompting ByteDance itself to issue a statement in an attempt to dispel some of the circulating claims.

In its statement, the tech giant confirmed that the sabotage did indeed take place in August of this year, and the intern in question has since been fired, with his university and relevant industry bodies – a.k.a. potential employers – being notified about the incident. At the same time, ByteDance denied losing tens of millions of dollars and using 8,000+ processors to train its AI models, calling these figures "seriously exaggerated."

Furthermore, the company claimed that the individual had been interning in the commercial technology team and had no internship experience in the AI ​​Lab, as was claimed by the intern himself on social media. As such, his sabotage is said to have only impacted the model training tasks of the commercial technology team's research project, and, according to ByteDance, "did not affect the formal commercial projects and online business, nor did it involve other businesses such as ByteDance's large models."

Mehaniq/Shutterstock

ByteDance's final supposed correction faced some backlash on social media, with one commenter challenging the distinction between the AI Lab and the commercial technology team and claiming the latter was previously part of the former. "In the past two years, the team's recruitment was written as AI Lab. He joined the team as an intern in 2021, and it might be the most advanced AI Lab," the comment read (via ArsTechnica).

AI Lab or not, the tech giant's statement at the very least acknowledged the disruption, and ByteDance deserves credit for not attempting to sweep it under the rug. The only thing that remains unknown is the intern's motive – is he an unsung hero of the Humans vs. Machines war, driven solely by his own ideals? Or could he be a borderline criminal paid by ByteDance's competitors to undermine the TikTok developer's AI initiatives? It's also possible, albeit the chance is incredibly slim, that the malicious code was added by accident, and the entire "sabotage" can be explained by trivial incompetence. Only the intern knows for sure.

What are your thoughts on the TikTok AI sabotage? What do you think is the most likely reason behind it? Share your thoughts down in the comments!

Don't forget to join our 80 Level Talent platform and our Telegram channel, follow us on InstagramTwitterLinkedInTikTok, and Reddit, where we share breakdowns, the latest news, awesome artworks, and more.

Join discussion

Comments 0

    You might also like

    We need your consent

    We use cookies on this website to make your browsing experience better. By using the site you agree to our use of cookies.Learn more