Samsung workers leaked sensitive information on ChatGPT

Share the joy

Samsung Semiconductor’s fab engineers used ChatGPT to swiftly patch mistakes in their source code after being given permission to do so.

In the process, they accidentally leaked private information like meeting minutes and statistics on the performance and yields of the fab.

The business now intends to create its own internal ChatGPT-like AI solution. However, the length of questions submitted to the site is currently restricted to 1024 bytes, according to Economist.

Samsung Semiconductor has so far identified three cases of ChatGPT use that resulted in data leakage. Three may not seem like much. But they all occurred over a period of 20 days. The situation is very troubling now.

In one instance, a Samsung Semiconductor employee uploaded the proprietary program’s source code to ChatGPT to address bugs. It effectively disclosed the application’s code to an external company’s artificial intelligence.

The second incident may have been much more concerning. Another employee requested optimization while entering test patterns intended to find damaged chips.

Test plans intended to find flaws must be private. With reduced costs, silicon test and verification processes can be sped up by streamlining these test sequences. It can reduce their number.

Another worker created a paper from a meeting recording using the Naver Clova app. The worker uploaded it to ChatGPT to create a presentation.

As a result of these efforts, Samsung has informed staff members about the risks associated with utilizing ChatGPT. Executives and staff at Samsung Electronics received information that data entered into ChatGPT is sent to and stored on other servers. It’s impossible for the business to retrieve it and raising the possibility of sensitive information leaking.

The open learning data component of ChatGPT has the potential to reveal private data to outside parties. This is unacceptable in the fiercely competitive semiconductor sector.

To avoid further mishaps, the corporation is already putting preventative measures in place. Emergency information protection measures are in place. It restricts access to ChatGPT on the corporate network if there is another incident.

The future of chip manufacturing will undoubtedly include technologies like generative AI and various AI-enabled electronic design automation techniques.

A spokeswoman for Samsung Electronics declined to confirm or reject the information. The information leakage incident was an internal concern.


Share the joy

Author: Francis Rey

Francis is a voracious reader and prolific writer. He has been writing about social media and technology for more than 10 years. During off hours, he relishes moments with his wife and daughter.

Share This Post On