Samsung finds chip-related data leak due to use of ChatGPT: Korean media

A ChatGPT logo is displayed at an office in Washington, D.C., U.S., March 15, 2023. /CFP

Samsung Electronics has found three incidents of leaked corporate information, some related to crucial semiconductors, South Korea's The Economist magazine reported.

The article, published on March 30, said that three incidents of information leakage had occurred in its device solutions business premises, or Samsung DS, which overseas the research and development of semiconductors.

The incidents were confirmed by the magazine as employees' input of corporate information into ChatGPT, the artificial intelligence chatbot developed by U.S. company OpenAI, was allowed at Samsung DS since March 11.

It added that information such as "equipment measurement and yield data were sent intact to U.S. companies" when employees were seeking code optimization from the chatbot, and another employee used it to generate meeting minutes.

The company had issued an internal circular warning about the misuse of ChatGPT, according to both the magazine and and a separate report by The Korea Times.

In an updated "Terms of Use" OpenAI claimed that it would not use input content from its Application Programming Interface (API) to train the model, but may collect non-API data to improve its services and users can opt out by filling a form.

A screenshot of OpenAI's "Terms of Use" from its website

More ban on ChatGPT due to lack of transparency

How data is being handled at OpenAI to train its AI model is still unclear and there's a growing concerns among companies about sensitive information being put into the chatbot.

ChatGPT was forced to be taken offline last month over a data breach involving users conversations' and information about subscriber payments.

The incident also triggered Italy to be the first Western country to curb ChatGPT over privacy concern, its Data Protection Authority announced last Friday.

A series of Wall Street firms, including JP Morgan Chase, Bank of America, and the Goldman Sacks Group, have restricted staff from using the ChatGPT chatbot over the past month according to Bloomberg reports.