Samsung Electronics to adopt own AI amid ChatGPT security concerns
| Samsung Electronics headquarters in Seoul / Korea Times file |
By Kim Hyun-bin
Samsung Electronics plans to introduce a customized artificial intelligence (AI) service for knowledge search, translation and summarization this year as concerns over the possible leakage of key technological information enabled by ChatGPT are rising.
The Device Solutions (DS) division, which oversees Samsung Electronics’ semiconductor business, plans to introduce its own large language model (LLM) that it assesses to be at a level higher than GPT-3.5 this year.
“Depending on how we utilize generative AI, we can bring about tremendous innovation in our work,” Samsung Electronics’ DS division head Kyung Kye-hyun said in a recent lecture for employees. “We are currently pursuing the adoption of customized AI through domestic specialized companies.”
Samsung Electronics DS division plans to open its basic services in December and launch professional search services incorporating company knowledge in February of next year.
Through this, the company plans to support employees’ work in automated responses for purchasing and expenses; specialized knowledge search for processes, design, and manufacturing; summarization of manufacturing and process data; translation; document creation; meeting transcription and summarization; market and company analysis; code generation and review; and voice of the customer (VOC) response.
“What I am currently doing is changing our organizational culture, restoring our weakened development competitiveness and meeting customers worldwide to draw the future together with Samsung,” Kyung said.
This move is in response to the high demand for generative AI such as ChatGPT in translation and document summarization tasks. However, there are also security concerns such as the risk of confidential information leaks due to the proliferation of generative AI. While the use of external generative AI like ChatGPT is restricted, the plan is to introduce customized AI internally to support efficient work processes.
In fact, the DS division has taken measures such as limiting the number of characters in the use of ChatGPT, following some cases of misuse, such as uploading internal mail content and the inputting of internal source code, discovered in March.
The Device Experience (DX) division, responsible for home appliances and smartphone businesses, has completely banned the use of generative AI through internal PCs since May.
Instead, the DX division has the policy to develop its own AI tools to support translation, document summarization and software development assistance. However, the specific development timeline and implementation schedule have not been finalized.
