OpenAI Suspends TikTok Owner ByteDance For Making Rival Bot Out Of Its Skeleton

Photos 269398084 © Awargula and 224914754 © Rafael Henrique | Dreamstime.com

GPT crickets! OpenAI, the creator of tools like ChatGPT and DALL-E, has suspended the account of ByteDance, the Chinese multinational conglomerate and parent company of TikTok. OpenAI took action following a report that ByteDance was utilizing OpenAI’s GPT AI technology to create a rival chatbot service, which would be in breach of the former’s terms of service.

ByteDance has been working on Project Seed, a secret and ambitious internal initiative focused on developing a large-language model. As The Verge learns from employees, however, the social media giant extensively used OpenAI’s GPT technology in almost every stage of Project Seed’s development. This purported usage seemingly conflicts with OpenAI’s policy that prohibits employing its AI platform to develop AI models that directly compete with its products and services.

ByteDance’s obtaining of access to OpenAI’s tools, reportedly through Microsoft, has also come under scrutiny. Microsoft, a principal investor and partner of OpenAI, has terms of service that align with those of OpenAI, making the alleged usage by ByteDance a potential violation.

Amid these developments, internal communications within ByteDance about its ‘Lark’ messaging service have also raised eyebrows. The company apparently sought to “whitewash” evidence of its utilization of OpenAI’s resources.

While ByteDance has dismissed the extent of its use of the OpenAI API as minimal, asserting that any related text was removed from the chatbot’s data set earlier this year, OpenAI has chosen to suspend its account pending further investigation.

A spokesperson for ByteDance reaffirms the initial use of OpenAI tools in the early stages of Project Seed with the publication, but maintains that the company had ceased using OpenAI-generated text.

This controversy emerges just as ByteDance received regulatory approval in China to launch its AI chatbot, Duobao, last August. The company reportedly instructed its team to discontinue using text generated by OpenAI around the same time.

In response to these allegations, Microsoft, a key player in this scenario, has emphasized its commitment to enforcing its code of conduct. The tech giant has procedures in place to identify misuse and is prepared to terminate access if any company violates its terms of service.

[via New York Post and DigiTimes, cover photo 269398084 © Awargula and 224914754 © Rafael Henrique | Dreamstime.com]