April 10, 2023
Employees at Samsung Semiconductor mistakenly revealed sensitive data while using OpenAI’s ChatGPT to assist them carry out tasks.
The employees were permitted to use the ChatGPT tool to troubleshoot programming issues, but they accidentally entered secret material such as the source code for a new software application, private meeting minutes, and hardware-specific details. Unfortunately, Samsung later discovered that three instances of staff personnel revealing personal information using ChatGPT occurred during a 20-day period.
As the ChatGPT platform retains user input data to improve its algorithms, Samsung’s sensitive information has now become OpenAI’s intellectual property. As a result, Samsung Semiconductor created its own AI system for internal usage by people, which limits prompts to a maximum of 1024 bytes.
According to Tom’s Hardware, one incident involves an employee who asked ChatGPT to improve test sequences, which is a highly secret operation that finds faults in chips. By shortening the processor testing and verification process, this innovation has the potential to save semiconductor firms time and money. In another case, a staff member used ChatGPT to convert meeting notes into a presentation that Samsung sought to keep private from other parties.
In response to these events, Samsung Electronics has issued a warning to its employees about the potential dangers of disclosing sensitive information. Because the information is now stored on OpenAI servers, it is impossible to retrieve or delete, leaving Samsung vulnerable to disastrous consequences in the highly competitive semiconductor industry.
The sources for this piece include an article in TechTimes.
