
Samsung Engineers Mistake ChatGPT for a Confidential Notepad
Ourtake
The chatbot doesn't keep secrets, but its training policy apparently does keep your prompts.
What happened
Soon after Samsung's semiconductor division allowed engineers to use ChatGPT, workers reportedly leaked confidential information to it at least three times, including sensitive database source code and a recorded meeting. Samsung reportedly limited prompt length, is investigating the three employees, and is building its own chatbot.
Sources
- economist.co.kr · [단독] 우려가 현실로…삼성전자, 챗GPT 빗장 풀자마자 ‘오남용’ 속출 (opens in a new tab)
- techradar.com · Samsung workers made a major error by using ChatGPT (opens in a new tab)
- engadget.com · Three Samsung employees reportedly leaked sensitive data to ChatGPT - Engadget (opens in a new tab)
- businessinsider.com · Samsung bans employees from using AI tools like ChatGPT and Google Bard after an accidental data leak, report says (opens in a new tab)