[Edaily Reporter Song Jae-min] As Samsung Electronics and SK Hynix move to adopt external generative artificial intelligence (AI), security concerns are emerging as a new topic of discussion. While the semiconductor industry, which had been cautious about using AI due to concerns over technology leaks, is shifting its focus toward productivity innovation, concerns about security risks remain due to the industry’s nature of handling national core technologies.

According to industry sources on the 15th, Samsung Electronics and SK Hynix are accelerating their expansion of external generative AI tools such as ChatGPT. Samsung Electronics has officially adopted ChatGPT, Gemini, and Claude in its Consumer Electronics (DX) division, and its Semiconductor (DS) division is also pushing to adopt ChatGPT and Gemini. SK Hynix is also considering the adoption of ChatGPT Enterprise and Microsoft (MS) Copilot.

This marks a stark shift from the atmosphere just a few years ago. Until recently, the semiconductor industry had been extremely cautious about using external generative AI due to concerns over the leakage of design data, process information, and customer data.

In fact, Samsung Electronics imposed a complete ban on the use of external generative AI in 2023 after it was revealed that source code had been entered into ChatGPT. Since then, the company has been utilizing its proprietary AI model, "Samsung Gauss," while SK Hynix has been operating an in-house AI service based on open-source technology.

The semiconductor industry is highly security-conscious due to the risk of technology leaks. In recent years, a series of incidents involving former Samsung Electronics and SK Hynix employees moving to Chinese semiconductor companies and leaking technology has heightened the industry’s vigilance.

Nevertheless, analysts suggest that both companies have moved to adopt external generative AI because AI utilization is no longer an option but a necessity. They note that a sense of urgency—driven by the intensifying competition in AI-based business innovation—has led them to act, as they cannot afford to fall behind.

Choi Byung-ho, a professor at the AI Research Institute at Korea University, said, “Security risks naturally exist,” adding, “Nevertheless, companies are adopting AI because they judge that the benefits of business process innovation are that significant.” He continued, “They will likely start by utilizing AI in areas unrelated to security or attempt to mitigate risks through various methods, such as encryption, access control, and the establishment of private environments.”

Companies are also taking action. Samsung plans to establish dedicated AI organizations across all its affiliates and strengthen its data control systems. Recently, the company has also expanded its recruitment of personnel for AI security management and the operation of AI security policies. This stems from the judgment that as the adoption of generative AI increases, security systems must be built alongside it.

However, experts point out that security concerns have not been completely resolved. Professor Choi said, “The level of risk can vary depending on how it is operated,” adding, “Finding a balance between expanding AI utilization and protecting technology will be an important challenge for companies going forward.”