Samsung is developing a private generative AI tool

A Samsung executive has confirmed rumors that the company is developing a ChatGPT-like generative AI tool. The confirmation came from co-CEO Kyung Kye-hyun during a recent internal meeting, anonymous industry sources told the Korean publication The Investor. Kyung oversees the tech giant’s chip business.

Samsung will keep its generative AI private

Unlike ChatGPT and Google Bard, Samsung’s generative AI service will be private to the company. The general public can’t access it. The Korean firm will only allow some of its employees to access the AI tool, which doesn’t yet have a name. The idea is to use AI to improve and speed up business processes wherever possible.

According to the new report, Samsung sees AI helping its business in nine key areas. Those include data analysis and information sourcing for chip production, translation of documents and meetings, transcribing audio files or meeting voices, cost management, and automated responses for product purchases.

Samsung previously tried to use ChatGPT to help its semiconductor division in business growth. However, it didn’t work out as intended. Employees inadvertently ended up leaking critical company information through the OpenAI-made tool. The Korean biggie has since banned the use of all external AI tools on official devices and internal networks.

However, Samsung does plan to incorporate ChatGPT in its products or services in some way next year. During a separate event at Yonsei University last Friday, Hyung described the tool as the “best intellect” in this space currently. He didn’t share further plans, though. The company is reportedly exploring new ways to contribute to the growing AI ecosystem beyond its internal goals.

“Depending on how we use generative AI, there can be tremendous innovation in what we do,” said Kyung. “We’re promoting the introduction of customized AI through a local specialized firm”. The Samsung co-CEO suggested potential applications of AI solutions in “wide bandwidth, high power efficiency and integration architecture from smartphones to data centers.”

Samsung’s new AI tool is still months away

It was rumored last week that Samsung could complete the work on the initial version of its new LLM (Large Language Model) as early as next month. However, the AI tool may not be ready for prime time until next year. The new report states that the company is aiming to launch the “basic service” in December this year. It may fully integrate the tool with its systems by February 2024.