Samsung Allows External Generative AI Use for DX Employees

Samsung Electronics will allow employees in its DX division to use external generative AI models starting in June, following a proof-of-concept with 2,500 staff that tested Gemini, ChatGPT, and Claude. The company is drafting operating policies and will restrict access to employees who complete security training, while continuing to develop its in-house model Samsung Gauss. The move aims to accelerate decision-making and improve productivity across product planning, marketing, and global operations.

Samsung Allows External Generative AI Use for DX Employees Reporting by Digital Today and BizChosun on May 26 says Samsung Electronics will permit employees in its DX division to use external generative AI models starting in June. Digital Today reports the company ran a proof-of-concept in April and May with 2,500 employees, testing Gemini, ChatGPT, and Claude. Both outlets say Samsung is creating detailed operating policies, will restrict access to staff who complete security training, and intends to run external services in parallel with its in-house model Samsung Gauss. Reporters frame the move as intended to speed decision-making and improve productivity across product planning, development, marketing, and multilingual global operations. What happened Reporting by BizChosun and Digital Today on May 26 says Samsung Electronics will introduce external generative AI services for employees in its DX finished goods division beginning in June. Reporting by Digital Today says the company executed a proof-of-concept in April and May with 2,500 employees and tested Gemini, ChatGPT, and Claude. Both outlets report Samsung is drafting detailed operating policies, targeting an official launch in June, and limiting access to employees who complete security training. Digital Today and BizChosun also report the company intends to keep developing its internal model Samsung Gauss while using external services in parallel. Technical details Reporting by BizChosun and Digital Today lists the intended enterprise use cases as insight generation during product and service planning, support for global marketing and communications, multilingual overseas business response, and market and customer data analysis. Digital Today reports that access to tested external services will be gated by completion of security training and that detailed policies will reflect job functions and organisational characteristics. Editorial analysis - technical context Enterprise pilots that combine internal models with external large-model services are a common pattern for organisations balancing in-house IP and rapid access to innovation. Companies running similar parallel tracks typically face engineering work to unify model outputs, manage prompt engineering across diverse backends, and implement consistent input-output sanitisation to limit data leakage risks. For practitioners, expect integration work to center on access control, prompt templates, secure connectors, and monitoring to make externally generated outputs auditable and reproducible. Context and significance Reporting in Digital Today frames the move as an attempt to absorb strengths of external models and close an "AI utilisation gap" with global competitors. Industry-pattern observations show large enterprises often pursue restricted, role-based rollouts when opening external AI access; those rollouts frequently prioritise non-sensitive workflows such as ideation, summarisation, and multilingual drafting before expanding into customer- or IP-sensitive tasks. For data teams, the significance is practical: once approved, external models become part of the toolchain practitioners must evaluate for accuracy, hallucination risk, privacy constraints, and MLOps integration. What to watch For practitioners: monitor the published operating policies for specifics on allowed query types, logging and retention rules, and approved connectors. Also watch whether access expands beyond the DX division, how Samsung Gauss development proceeds in parallel, and which commercial providers OpenAI, Google, Anthropic, etc. are adopted long term. Observers should track telemetry and governance metrics-adoption rates, incident reports, and red-team findings-to assess how securely external models are being operationalised. Scoring Rationale The story is notable because Samsung is a major hardware and consumer-electronics firm; opening internal teams to external LLMs affects product planning and global operations. The move is important for practitioners who will integrate, govern, and monitor external model use alongside an internal model. Practice interview problems based on real data 1,500+ SQL & Python problems across 15 industry datasets — the exact type of data you work with. Try 250 free problems /problems