Sunday, 30 November 2025
25.7 C
Singapore
13.3 C
Thailand
24 C
Indonesia
27.7 C
Philippines

OpenAI halts ByteDance’s access for alleged misuse of AI tech

OpenAI suspends ByteDance's account over alleged misuse of AI technology in developing its chatbot amidst policy violations and regulatory scrutiny concerns.

ByteDance, the company behind the globally popular app TikTok, has reportedly been using OpenAI’s technology to enhance its own artificial intelligence model, as per a report by the Verge. This move is frowned upon in the AI community and breaches OpenAI’s terms of service.

Under these terms, customers are prohibited from creating AI models rivalling OpenAI’s offerings. They are also restricted from extracting data through any means other than the approved Application Programming Interfaces (APIs). These APIs enable developers to utilise GPT to build their own applications.

Despite these regulations, ByteDance allegedly continued employing the API to train and assess its model, Project Seed. Internal communications within ByteDance, viewed by the Verge, indicate instructions to staff to conceal their activities using “data desensitisation” methods.

ByteDance reportedly halted using the API for Project Seed development when its chatbot, Doubao, received regulatory approval in China. However, the Verge’s report suggests that the API was still being used to evaluate Doubao’s performance.

OpenAI responds with account suspension

Reacting to these allegations, OpenAI confirmed on Friday that it had suspended ByteDance’s account for violating its terms of service. “All API customers must adhere to our usage policies to ensure that our technology is used for good. While ByteDance’s use of our API was minimal, we have suspended their account while we further investigate,” said OpenAI spokesperson Niko Felix.

If ByteDance’s use is found to be against OpenAI’s policies, the company could be asked to make necessary adjustments or face account termination.

ByteDance has denied any misconduct in a statement to Business Insider, asserting that it holds a licence from Microsoft to use GPT’s APIs. The company clarified that it employs GPT for products in markets outside China, but uses its in-house model for Doubao, which is exclusive to China.

ByteDance’s expansion into AI and regulatory scrutiny

ByteDance is currently broadening its horizons, venturing into the AI market. The South China Morning Post recently reported that the company is developing a chatbot-building tool, positioning itself as a competitor to OpenAI’s customisable ChatGPT bots.

In the past year, US regulators and lawmakers have been closely monitoring ByteDance, expressing concerns that user data from TikTok might be shared with the Chinese government.

Hot this week

The forgotten battle royale that ended a studio still deserved more than a one-month run

A look back at Radical Heights, the short-lived battle royale that showed promise but shut down after just one month.

Warner Music ends lawsuit against Suno after reaching new licensing agreement

Warner Music ends its lawsuit against Suno after securing a licensing deal that gives artists opt-in control over AI-generated music.

Singapore sees surge in ransomware attacks during holidays, Semperis study finds

A new Semperis study shows 59% of ransomware attacks in Singapore occur during holidays, driven by reduced staffing and major corporate events.

Chrome tests new privacy feature to limit precise location sharing on Android

Chrome for Android tests a new privacy feature that lets websites access only approximate location data instead of precise GPS information.

Valve offers strongest hint yet on expected Steam Machine pricing

Valve hints that the Steam Machine may be priced close to a similarly powerful DIY PC, but external factors keep final costs uncertain.

DeepSeek launches open AI model achieving gold-level scores at the Maths Olympiad

DeepSeek launches Math-V2, the first open AI model to achieve gold-level scores at the International Mathematical Olympiad.

AI browsers vulnerable to covert hacks using simple URL fragments, experts warn

Experts warn AI browsers can be hacked with hidden URL fragments, posing risks invisible to traditional security measures.

Slop Evader filters out AI content to restore pre-ChatGPT internet

Slop Evader filters AI-generated content online, restoring pre-ChatGPT search results for a more human web.

Lara Croft becomes gaming’s best-selling heroine amid new Tomb Raider rumours

Lara Croft becomes gaming’s best-selling heroine as new Tomb Raider rumours fuel excitement.

Related Articles

Popular Categories