JAL Launches Humanoid Robot Ground Staff Trial at Haneda Airport: G1 and Walker E to Handle Cargo and Cabin Cleaning.

April 20, 2026 · International Civil Aviation and Artificial Intelligence · Compiled from JAL Group press releases and related reports on GMO AI & Robotics.

Overview: According to a press release from Japan Airlines (JAL) Group, JAL Ground Service (JGS), which is responsible for ground handling at JAL Group's major domestic airports, will collaborate with GMO AI & Robotics (GMO AIR), a subsidiary of GMO Internet Group, to conduct a verification test of humanoid robot applications at Haneda Airport starting in May 2026. The press release states that this will be Japan's first verification test of humanoid robot applications in an airport setting, with the focus on tasks such as cargo handling and cabin cleaning.

1. Content of the experiment: Who will conduct it, and what will they do?

According to the press release outlining the division of labor, JGS will handle ground support operations at JAL Group's major domestic airports, including aircraft towing, baggage and cargo loading/unloading, etc. GMO AIR will be responsible for promoting the social implementation of AI and robotics within the GMO Internet Group. Under the banner of "Japan's first airport humanoid robot application verification," the two companies will launch a trial starting in May 2026. According to related reports, the trial will utilize humanoid robots such as Unitree's G1 and UBTech's Walker E, and will involve tasks such as cargo handling and cabin cleaning.

Choosing cargo handling and cabin cleaning as initial areas of focus makes sense because these tasks are highly repetitive, physically demanding, and involve relatively structured workspaces, which makes it easier to define the robot's operational boundaries and safety protocols. Higher-risk tasks, such as aircraft towing, are expected to be automated in later phases.

2. Why humanoid robots? Addressing labor shortages in ground handling.

From a broader perspective, airports have long faced challenges in recruiting ground staff, particularly for tasks such as apron operations and cabin cleaning during nighttime hours. Japan's shrinking working-age population has exacerbated this issue. The appeal of humanoid robots lies in the fact that they do not require significant modifications to existing infrastructure: boarding gates, cargo bay doors, and cabin aisles are all designed according to human dimensions, meaning that humanoid robots can theoretically directly enter and operate within spaces designed for humans.

However, there is still a gap between verification and actual deployment. According to the framework disclosed in the press release, this round is a demonstration experiment. Questions such as whether the robots can operate stably for a long period, how to ensure that malfunctions do not affect flight operations, and how to establish safety rules for human-robot collaboration all need to be answered during the trial period.

3. JAL's AI Strategy: Another Approach Beyond Robots

Humanoid robots are not the only AI initiative at JAL. According to industry reports, JAL is also introducing Microsoft's Phi-4 small language model to enhance its ability to process safety reports, enabling faster categorization and analysis of safety information submitted from the front lines. It can be seen that JAL's current AI applications are more focused on back-end operations such as safety, operations, and ground services, rather than passenger-facing interfaces. This contrasts with some airlines in Europe and the United States that are focusing on AI-powered customer service and search functions, and it aligns with JAL's operational priorities under its "Safety Charter" framework.

4. Brief commentary.

The key technological aspect of this event lies in bridging the gap "from demonstration to deployment": both the U-Tree G1 and the UBTECH Walker E are commercially available models, and the hardware itself is not scarce. What is scarce is how airlines, airports, and robot operators can divide tasks, share risks, and assign responsibilities. Japan's approach, treating this as a national-level "first verification" project, makes it worth tracking not how fast the robots move, but rather the operational standards and safety protocols that emerge from the trial. Once this mechanism is established and proven effective, it is likely that it can be replicated in other major airports more quickly than the robot technology itself can be improved.

Source: This article is compiled from JAL Group’s official press release (April 2026), related reports on GMO AI & Robotics, and publicly available industry information. The data is based on the original reports and is accurate as of the date of publication (April 20, 2026). Compiled from publicly available information and for reference only.