Japan Deploys AI Robots in Wet Labs to Tackle Science Talent Crisis

Japan has long faced a paradox at the frontier of scientific research: a country renowned for precision manufacturing and engineering excellence is quietly running out of the scientists and clinicians needed to keep pace with the demands of modern biomedical research. The solution now being deployed in laboratories across the country looks less like traditional automation and more like something out of a robotics research paper—AI-driven humanoid machines capable of independently executing complex experimental protocols around the clock.

The Institute of Science Tokyo, formed from the 2024 merger of Tokyo Institute of Technology and Tokyo Medical and Dental University, announced the opening of a Robotics Innovation Center featuring Maholo humanoid robots developed jointly with Japan’s National Institute of Advanced Industrial Science and Technology (AIST) and industrial robotics giant Yaskawa Electric. The lab can run up to 1,000 distinct experiments 24 hours a day, 7 days a week.

Why Wet Labs Are the Bottleneck

Wet lab work—the hands-on manipulation of biological samples, reagents, pipettes, and culture dishes that underpins most biomedical research—has resisted automation longer than almost any other scientific task. It requires the kind of fine motor precision, contextual judgment, and real-time adaptation that traditional industrial robots handle poorly. A robotic arm optimized for repetitive factory assembly cannot easily pipette a microliter of solution into a 96-well plate without spilling, or recognize when a culture sample has been contaminated and needs to be discarded.

Humanoid robots change the equation. Designed to operate in environments built for human hands, they can use the same equipment that researchers do—benchtops, centrifuges, microscopes, incubators—without requiring specialized laboratory redesign. Maholo, developed over nearly a decade through the AIST-Yaskawa collaboration, can execute protocols ranging from sample preparation to cell culture maintenance with sufficient reliability to run continuously without human supervision.

Keiichi Nakayama, the center’s director, estimated the system could speed research “by 10 to 100 times”—a range that reflects the inherent variability of experimental biology, where some protocols benefit more from 24/7 operation than others. Even at the conservative end, a tenfold acceleration of the experimental iteration cycle is transformative: it compresses the timeline from hypothesis to preliminary data from months to weeks.

AI for Cancer Screening: Addressing the Cytologist Shortage

Wet-lab automation is only one front of Japan’s AI-in-medicine push. A parallel effort is deploying AI to analyze cell images for cancer screening, targeting a structural shortage in trained cytologists—specialists who manually examine stained cell samples under microscopes to identify malignancies.

Japan’s aging population means cancer screening demand is rising while the supply of trained cytologists has not kept pace. AI systems trained on large libraries of annotated cell images can flag suspicious samples for human review, functionally multiplying the throughput of existing cytology departments without requiring an equivalent increase in trained staff. By reducing the proportion of negative samples that require human expert review, these systems allow cytologists to focus their time on the genuinely ambiguous cases where human judgment adds the most value.

AI Medical Services (AIM), a Tokyo-based healthtech startup, has developed endoscopy AI models trained on more than 200,000 high-resolution videos from over 100 institutions—one of the largest annotated medical imaging datasets assembled in Japan. The company’s models detect polyps and other gastrointestinal abnormalities in real time during endoscopic procedures, and are now deployed in hundreds of hospitals across the country.

The Tohoku Medicinal Hub

A third pillar of Japan’s AI-in-medicine strategy is the Medicinal Hub at Tohoku University, which is taking a systemic rather than tool-specific approach. The Hub functions as a node linking clinicians, AI researchers, and health-technology firms around a shared repository of patient data, experimental results, and computational infrastructure.

The logic is that many of the barriers to AI deployment in medicine are not technical but organizational: the data sits in hospital systems, the researchers sit in universities, and the companies that could productively use both are neither. By co-locating all three in a structured collaboration environment, the Hub aims to reduce the friction that has historically kept Japan’s considerable biomedical data assets from being productively applied.

The Budget Behind the Push

The scale of Japan’s commitment to physical AI robotics reflects both the urgency of the talent crisis and the government’s read on where competitive advantage in the AI era will be won. Japan’s fiscal year 2026 budget allocates over 387.3 billion yen—approximately $2.6 billion at current exchange rates—toward the development of a multimodal infrastructure model specifically geared toward AI robots and physical AI systems.

This is not simply a research subsidy. It is a bet that Japan’s existing strengths in precision manufacturing, robotics hardware, and industrial systems integration give it a structural advantage in physical AI that it lacks in the large-language-model arms race dominated by American and Chinese players. Rather than competing directly on frontier model capability, Japan is placing its chips on the application layer—the deployment of AI-driven physical systems in hospitals, factories, and research institutions where the country’s engineering culture is already a proven asset.

The Broader Significance

What Japan is building in its wet labs and cancer screening clinics is an early glimpse of a pattern that will play out across every country facing the same fundamental constraint: the number of expert humans needed to address modern scientific and medical challenges exceeds what any training system can produce in time.

The solution is not to replace human experts but to amplify what they can do—automating the repeatable, quantitative, time-consuming parts of their work so that human attention can focus on the interpretive, relational, and creative dimensions that remain genuinely hard to automate. Japan’s AI-in-medicine push is less about replacing scientists and doctors than about making it possible for fewer of them to accomplish what previously required many more.

The combination of Maholo robots running experiments while oncologists sleep, AI cytology tools pre-screening the negative samples so specialists only see the ambiguous ones, and institutional hubs connecting data to researchers to companies represents a coherent national strategy for surviving a demographic crunch that Japan faces earlier than most—and that most of the world will face eventually.