South Korea is building a new playbook for training robots: capture how humans actually work, then teach machines to replicate those movements with precision. A Seoul-based artificial intelligence startup called RLWRLD (pronounced "real world") is wiring up hospitality workers, logistics staff, and convenience store employees with body cameras to harvest their expertise into a database that will eventually power AI-driven robots across industries.

The approach reflects a fundamental shift in how companies think about robot training. Rather than programming robots for specific, repetitive tasks, RLWRLD is collecting granular data on how skilled humans perform nuanced work, from folding napkins at five-star hotels to organizing food displays at convenience stores. This human-centered training method could give South Korea a competitive edge in the global race to build humanoid robots that can work alongside people in real-world environments.

Why Are Companies Capturing Worker Movements?

David Park, a food and beverage worker at the Lotte Hotel Seoul, has become part of an unusual experiment. Strapped with body cameras on his head, chest, and hands, Park performs his daily tasks while engineers record every motion. "I've been doing this about once a month," Park explained, noting that the cameras on his hands felt tight as he folded banquet napkins and polished glassware.

The data Park and other workers generate is converted into machine-readable information that trains test robots. RLWRLD's engineers add another layer by repeating the same tasks while wearing cameras, virtual reality headsets, and motion-tracking gloves. This dual-capture approach captures fine details such as joint angles and the amount of force applied, allowing robots to learn not just what to do, but how to do it with human-like precision.

"Capturing motion data in real-world settings is extremely important and the quality of that data matters greatly," said Hyemin Cho, who handles business strategies at RLWRLD.

Hyemin Cho, Business Strategy Lead at RLWRLD

This strategy mirrors how large language models like ChatGPT train on vast amounts of text data. Just as AI chatbots require extensive written examples to understand language, AI robots need extensive data on human action to handle advanced physical tasks. South Korea believes it has a natural advantage here: a deep base of skilled workers in manufacturing, hospitality, and logistics who can provide high-quality training data.

How Is RLWRLD Building Robot Intelligence From Human Skills?

• Multi-Source Data Collection: RLWRLD partners with major companies including the Lotte Hotel Seoul, logistics firm CJ, and Japanese convenience store chain Lawson to capture worker movements across different industries and skill levels.
• Motion Capture and Conversion: Worker footage is converted into machine-readable data, then engineers repeat the tasks while wearing motion-tracking equipment to capture precise details like joint angles and force applied during each movement.
• Robot Training and Refinement: The collected data trains test robots, often guided by RLWRLD "pilots" using wearable devices, allowing the AI system to learn both the sequence of movements and the subtle variations that make tasks successful.

One of RLWRLD's labs occupies a cluttered 34th-floor suite at the Lotte Hotel, filled with wires, computing equipment, and infrared laser readers positioned in the corners. Beneath a chandelier, a wheeled robot with black, humanlike metal hands moves back and forth, demonstrating the results of this training process. During recent demonstrations, the robot carefully lifted and placed cups at a minibar, while more advanced test footage shows a humanoid opening a box, placing a computer mouse inside, closing it, and setting it on a conveyor belt.

The focus on five-fingered hands sets RLWRLD apart from many competitors. Most robots, including Boston Dynamics' Atlas, use task-specific hands with two or three-fingered "grippers" designed for factory work. RLWRLD is among a smaller group developing AI for five-fingered hands that mimic human touch, a capability that could prove crucial as robots move into homes where closer interaction with humans will be required.

When Will These Robots Actually Start Working?

The timeline is becoming concrete. RLWRLD expects industrial AI robots to be deployed at scale around 2028, a timeline shared by major businesses in the region. Hyundai Motor plans to introduce humanoids built by its robotics unit, Boston Dynamics, at its global factories starting with its Georgia plant in 2028. Samsung Electronics plans to convert all manufacturing sites into "AI-driven factories" by 2030, with humanoids and task-specific robots across production lines.

The Lotte Hotel itself is planning for robot deployment. Although current humanoids would need several hours to clean a guest room that human workers finish in about 40 minutes, the hotel hopes robots will be ready for cleaning and other behind-the-scenes tasks by 2029. The company also plans to launch robot rental services for the hospitality and other service industries, with potential expansion to homes.

"If you look at the entire process of preparing for an event in back-of-house areas, we think humanoids might be able to take over about 30% to 40% of that workload. It will be difficult for them to replace the remaining 50%, 60% and 70%, which involves actual human-to-human interaction," said a Lotte Hotel representative.

Lotte Hotel Executive, Source 1

South Korea's government is also backing this effort. Last month, the government announced a $33 million project to capture the "instinctive know-how and skills" of "master technicians" into a database for AI-powered manufacturing, hoping robots will boost productivity and offset an aging, shrinking workforce.

Why Does South Korea Think It Can Win This Race?

South Korea faces stiff competition from U.S. tech giants like Tesla and a flood of Chinese firms pouring billions into humanoids and other AI robots. However, South Korean companies and policymakers believe they have a distinct advantage: a highly developed manufacturing sector with a deep pool of skilled workers who can provide high-quality training data.

The country struggled to compete in the chatbot market, where English language proficiency gives U.S. firms major advantages. But physical AI, the field of machines equipped with AI and sensors that can perceive, decide, and act in real-world environments, plays to South Korea's strengths. The nation's expertise in semiconductors, manufacturing, and robotics creates a natural foundation for building and training humanoid systems.

"South Korea has a highly developed manufacturing sector and the focus is squarely on humanoids tailored specifically for those industries," said Billy Choi, a professor at Korea University's center for Human-Inspired AI Research.

Billy Choi, Professor at Korea University's Center for Human-Inspired AI Research

However, this push has unsettled labor groups. After Hyundai's union warned in January that robots could trigger an "employment shock," President Lee Jae Myung issued a rare rebuke, describing AI as an unstoppable "massive cart" and calling for unionists to adapt to changes "coming faster than expected". Labor advocates worry that widespread robot deployments could hollow out the skilled workforce that South Korea is now counting on for its AI transition.

The stakes are high. South Korea's ability to capture and replicate human expertise through AI could determine whether it becomes a global leader in physical AI, or whether it gets left behind by faster-moving competitors in the United States and China. The workers being filmed in hotels and warehouses today are essentially training the robots that will reshape their industries tomorrow.