Source: Quantum Bit

Image Source: Generated by Wujie AI

ByteDance is embroiled in a controversy over large models.

According to The Verge:

ByteDance has been secretly using OpenAI's technology to develop its own large language model (LLM).

Shortly after this news was disclosed, The Verge further reported that OpenAI has suspended ByteDance's account.

Specifically, a statement released by OpenAI spokesperson Niko Felix is as follows:

Although ByteDance's usage of our API is minimal, we have suspended their account and will conduct further investigation.
If we find that their usage is not in compliance with the rules, we will require them to make necessary changes or terminate their account.

The "rules" mentioned here refer to a specific provision in OpenAI's terms of service, which states that the model capabilities provided by OpenAI are not allowed to be used to "develop any AI model that competes with its products and services."

It is understood that ByteDance accessed OpenAI through a purchase made by Microsoft, but Microsoft has also established a policy similar to OpenAI's.

The Verge stated that they are seeking further consultation with Microsoft to determine whether they will also take the same action to suspend ByteDance's account as OpenAI did.

So, what exactly is the plagiarism controversy this time?

Exposure of Internal Documents

According to The Verge, the evidence comes from an internal document of ByteDance—chat records of the overseas version of the Lark app.

This document indicates that ByteDance relied on OpenAI's API for almost every development stage in the "Project Seed" large language model project, including model training and evaluation.

"Project Seed" was launched about a year ago and is currently mainly developing two products: one is Doubao, which has already been launched in China; the other is a chatbot platform for commercial users, which is currently under development.

It is claimed that employees involved in "Project Seed" were well aware of the consequences of excessive reliance on OpenAI's API, so they began discussing how to embellish evidence through "data desensitization".

As a result, there were often situations where employees reached the maximum access limit of the OpenAI API.

More specifically, ByteDance used OpenAI's technology mainly in the early stages of "Project Seed."

According to The Verge, based on the internal document, ByteDance issued a command to "stop using text generated by GPT at any stage of model development" a few months ago.

However, it was also at this time that ByteDance released its own large language model Doubao.

But The Verge stated that even at this time, ByteDance continued to violate the rules:

ByteDance continued to use the API in violation of OpenAI and Microsoft's terms of service, including evaluating the performance of the model behind Doubao.

And also stated that a person with first-hand information about ByteDance's internal situation pointed out:

They said they wanted to make sure everything was legal, but in reality, they just didn't want to get caught.

ByteDance's Response

After The Verge published this report, ByteDance spokesperson Jodi Seth made the following response:

Data generated by GPT was used to annotate models in the early development of "Project Seed" and was removed from ByteDance's training data around the middle of this year.
ByteDance has been authorized by Microsoft to use the GPT API.
We use GPT to support our products in non-Chinese markets; but in the Chinese market, we use our self-developed model to support Doubao.

On the Microsoft side, spokesperson Frank Shaw stated:

Microsoft AI solutions such as Azure OpenAI services are part of our limited access framework, meaning all customers must apply for and obtain approval from Microsoft.
We have also established standards and provide resources to help customers use these technologies responsibly and in compliance with our terms of service.
We have processes to detect abuse and will stop access for companies found to violate the code of conduct.

Quantum Bit also reached out to ByteDance for comment, but ByteDance has not yet made a formal response.

We will continue to follow the progress of this event in the comments section.

Reference links:
[1]https://www.theverge.com/2023/12/15/24003151/bytedance-china-openai-microsoft-competitor-llm
[2]https://openai.com/policies/business-terms

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。