Samsung engineers leaked sensitive data using ChatGPT

A report published by The Economist last week revealed this Confidential data from Samsung Semiconductor has been leaked thanks to the use of ChatGPT by engineers. The AI ​​tool was launched by the company due to the fact that it can help with mathematical calculations and in solving some programming problems.

but, It ended up bringing more headaches than benefits to the manufacturer, which had to deal with three cases of abuse in just 20 days. According to the report, the leaked information included measurement data and equipment results, which were impeccably transmitted to “US companies”.

To prevent similar situations from occurring in the future, Samsung has decided that it will develop an alternative to ChatGPT. In the meantime, the company has decided to limit the size of questions that can be sent to the AI ​​to 1024 bytes for the time being.

Violations were considered serious

According to The Economist, the first breach occurred when The engineer used the tool to analyze the complete source code of proprietary software. In the second case, an employee introduced test patterns for a new chip into the system, and asked ChatGPT to help him make improvements to it.

The third hack involved an employee who used the Naver Clova app to convert a meeting recording into a text document. immediately, It has been entered into the AI ​​tool databasewhich was used to turn your content into a presentation.

In response to this situation, Samsung Electronics made a presentation to its executives and employees alerting them to the fact that everything entered into ChatGPT is transmitted and stored on external servers. The company warns that in the event of a new security breach, You may have to completely cut off access to technology in the future.

source: tom devices