Samsung is making its own ChatGPT-like AI to avoid company leaks

Samsung recently learned the dangers of using ChatGPT within a corporate environment the hard way. Bits of the company's sensitive semiconductor-related information leaked through ChatGPT when a few employees tried to use the tool to ease their work, without realizing that the information and blocks of code they were sharing with the generative AI would become a part of ChatGPT and be stored on remote servers outside of Samsung's grasp.

Since then, Samsung disallowed its employees to use ChatGPT, but reportedly, the company won't give up on the idea of assisting its workforce through the power of generative AI. It is reported that the company has partnered with Naver Corp. to jointly develop an AI platform specifically and exclusively for Samsung corporate users. (via The Korea Economic Daily)

Built using sensitive semiconductor information and company data

Samsung's generative AI won't be open like ChatGPT but exclusive to its employees across the Device Solutions Division. And later, after field tests, the tool may become available to other employees of other Samsung branches, such as the Device eXperience division, which is responsible for mobile phones, home appliances, and so on.

Because Samsung's in-house AI won't leave the company's servers and will be exclusive to its employees, it may be tailored to help the company better than ChatGPT ever could.

For example, Samsung will supposedly share sensitive semiconductor data with Naver, who will then implement this information into the generative AI. Doing so will allow Samsung employees to use the AI's true potential without worrying that semiconductor-related source code might leak into the public cloud space.

Better yet, this in-house AI chatbot, co-developed by Samsung and Naver, will understand the Korean language better than any other generative AI.