Sensitive Samsung data leaked on ChatGPT by employees

Sensitive Samsung data leaked on ChatGPT by employees

Employees of the South Korean company unwittingly leaked highly confidential data using the chatbot. Something to encourage caution in the use of this artificial intelligence.

From our correspondent in Seoul,

Whether ChatGPT revealed secret data from Samsung Electronics, it’s simply because employees of the South Korean firm, gave him this information… ChatGPT is a conversational agent, it continues to improve through exchanges that he talks to users. That is to say, he refines his knowledge thanks to the conversations and requests made to him. Everything transmitted to it is transformed into learning data, which makes it possible to continuously improve the performance of artificial intelligence. But the consequence is that what we share with ChatGPT, we potentially also share with other users.

Leak in semiconductors

Employees of the company have, it seems, disclosed confidential information on three occasions. A first time, a person who had detected a problem in the execution of a program for downloading a database copied the entire source code into ChatGPT so that we could offer him a solution. Then, another person wanted to “optimize” some code to make it more efficient, so they grabbed all the information for the AI ​​to do it. Finally, the third person converted the minutes of a meeting recorded on his phone and forwarded it to ChatGPT asking him to give a presentation.

► To listen also: [Décryptage] Chat GPT, the robot that fascinates and worries

Three errors which notably led to the leak of certain data in the key sector of semiconductors. She quickly decided to limit download capacities in ChatGPT to 1,024 bytes per person and launched an investigation into the various leaks.

Samsung wants to create its own AI service

The company plans to create its own artificial intelligence to simplify the work of its employees without revealing confidential data. Online seminars have been organized for company executives to enable them to adapt to technological developments.

The paradox is that Samsung had already considered that the use of artificial intelligence could generate a potential data leak, because only some of the employees were authorized to use ChatGPT to better understand its potential. It seems impossible for Samsung to recover the leaked data, because in the user guide of artificial intelligence, it was clearly stipulated that sensitive information should not be revealed in the conversations.

rf-5-general