TikTok owner ByteDance has said it sacked an intern after they disrupted the company’s AI model training, but denied reports that the incident caused tens of millions in damages, as had been rumoured.

Posting on the company’s news aggregation platform, Toutiao, ByteDance said that the doctoral intern had been part of the commercialisation wing of the business, before being fired in August for ‘maliciously interfering’ with its AI after apparently being dissatisfied with the allocation of company resources.

Rumours had been circulating on Chinese social media that the episode had left a $10m (£7.7m) dent in the company’s finances, and that the unnamed intern had impacted the work of over 8,000 GPUs. According to media reports, the intern had also claimed to have inserted ‘malicious code’ into the AI model.

In its statement, ByteDance tried to downplay the incident and counter those claims, saying that “The online rumour that more than 8,000 GPUs were involved and the loss of tens of millions of dollars was seriously exaggerated.

“The person involved has been interning in the commercial technology team and has no internship experience in the AI ​​Lab. His social media profile and some media reports are incorrect.”

The Chinese tech giant confirmed that the intern hadn’t caused any lasting damage, and that its core business and AI development were still on track.

Already one of the world’s biggest tech companies thanks to the ubiquitous use of TikTok across the world, ByteDance’s Doubao LLM model was released in May this year,with the company claiming the AI chatbot already has 26 million active users on mobile and PC in China.

Recommended reading

• TikTok To Flag All AI-Generated Images
• TikTok Hit With Hefty Fine For Providing Inaccurate Data to Ofcom
• TikTok Handed €345M Fine for Handling of Children’s Data

Earlier this year, the company’s Doubao chatbot surpassed its main competitor Baidu’s ‘Ernie’ in the race to create a Chinese alternative to OpenAI’s ChatGPT. 

This month, ByteDance made headlines after cutting hundreds of jobs across its global workforce, as TikTok moves to automate more of its content moderation operations to rely more heavily on AI.

According to the company, 80% of the social media platform’s content is removed by automation, with the firm looking to invest $2 billion (£1.5 billion) in trust and safety across their global operations this year.