Samsung Leaks Confidential Information by using ChatGPT
April 15, 2023
Samsung Leaks Confidential Information by using ChatGPT? ChatGPT became a headache for Samsung. The fame of the artificial intelligence-powered chatbot led to several organizations beginning to integrate it into their workflows.
One of them has been Samsung Semiconductor. Where engineers used it to fix bugs in their product code without knowing that they were leaking confidential information.According to an Economist report, Samsung employees leaked sensitive content in 3 situations, some of them have gotten source code.
The incidents occurred within twenty days, after the big South Korean allowed the use of ChatGPT in its semiconductor separation. In consensus with the newspaper, the workers sought to take advantage of 2 key properties of the chatbot without being aware that the data would culminate in OpenAI.
The first leak has been from an engineer in charge of an internal database program. The employee was found with a compiler error and it was easy for him to mimic and paste the source code into ChatGPT to figure out the solution.
Although many turn to the chatbot to solve programming problems, OpenAI tells users not to enter sensitive information. The second leak was granted after another employee entered a code used to detect bad chips and requested a ChatGPT to optimize it.
This leak could be expected as the most sensitive due to secrecy in the construction processes at Samsung. Finally, a third worker entered a file with the details of a meeting for the AI to produce a presentation.
Faced with the leaks, Samsung applied emergency measures and imposed an upload limit of 1,024 bytes to the questions that are made in ChatGPT. The company will contribute to make an investigation and will punish the perpetrators if it is essential.
It is possible that the leak of confidential information from Samsung is not the first to receive ChatGPT. Since its rise to fame, a huge number of users have turned to OpenAI’s AI (artificial intelligence) to ask questions, fix issues, and improve content.
Some programmers enter source code while waiting for artificial intelligence to identify errors or optimize the program. What few people know is that all the information they enter into ChatGPT goes to OpenAI’s servers and is used to practice the model. This has already set off alarm bells in several companies, which restrict or ban their employees from using the chatbot.
Amazon has already sent ads to its workers, in which Verizon and JP Morgan Chase chose to ban the use due to fear of a leak. Samsung’s situation is fundamental, because the confidential data came out of one of its most relevant divisions.
Samsung Semiconductor is the delegate of generating the processors, image sensors, memory and storage resolutions. All of this used by both its products and other third-party devices. A boss to detect defective chips needs a particular shield in an industry as competitive as the semiconductor industry.
According to the report, Samsung will finally wait for the inquiries to make a choice. The company makes it possible for certain of its employees to use ChatGPT to familiarize themselves with technology. Though it may ban it if leaks persist. Divisions such as mobile, user experience and smart home are not authorized to use in their work groups.