Interview

What led me to join ITLAB was a traffic accident in 2000, when I was struck by a taxi. That experience made me strongly believe that cars must become safer, and it was this conviction that motivated me to join ITLAB.

For example, many modern vehicles are equipped with Advanced Driver Assistance Systems (ADAS), which include functions that can follow the vehicle ahead on highways. However, these systems do not automatically guarantee safety. In reality, there have been accidents where drivers over-relied on such systems, became inattentive, and crashed into road construction sites. Driver assistance functions do not always operate perfectly, and there are situations in which humans—who can perceive a wider range of their surroundings—are better at detecting potential hazards. Even as vehicles become more advanced, they are still driven by humans. Therefore, achieving zero accidents requires us to consider how to encourage safe driving behavior and how to guide human behavior in a safer direction.

Currently, our research focuses mainly on two themes. One is supporting drivers’ safety checks at intersections and stop signs on surface streets, where accidents frequently occur. The other is preventing drowsiness and inattentive driving caused by monotonous driving on highways.

In both cases, safety can only be achieved not through sensing alone, but through appropriate engagement with the driver. Interactions between drivers and vehicles are critically important for realizing zero accidents through a unified human–vehicle system. During the 2010s, we worked on technologies that sense the workload imposed on drivers and control the amount of information and available operations presented by in-vehicle systems. These technologies were subsequently implemented in real vehicles.

Since fiscal year 2025, we have stepped up our efforts to prevent inattentive driving. Going forward, it will be important to address not only obviously inappropriate driver states—such as looking away or falling asleep—but also inattentive states that may appear unproblematic at first glance, where drivers are seated properly and looking ahead. As ADAS becomes more widespread, vehicles will take over some driving operations, leaving drivers in a state where they must remain visually attentive even when not actively driving. This raises concerns that, despite appearing safe externally, drivers may cognitively drift into a state of inattentiveness. To address this, it is necessary to assess not only the driver, but also external environments and vehicle states in an integrated manner.

My primary focus is on helping elderly drivers eliminate bad habits that they have acquired over many years of driving after obtaining their licenses. My approach spans multiple disciplines, including psychology, social infrastructure, and physiology.
In my previous job, I was involved in the development of car navigation systems and initially worked with ITLAB as an external partner. I eventually joined ITLAB as a researcher after connecting with the former president. Car navigation systems represent an interface between humans and vehicles, and they incorporate extensive behavioral science research in areas such as user interface design and the timing and manner of information presentation. Building on that foundation, my work began with research into AI-based alert systems linked to navigation and positioning functions, and has since expanded into fields such as neuroscience and physiology.

When thinking about behavioral change, we often focus on directly influencing people. However, there are also cases where individuals interpret situations on their own and change their behavior accordingly. One example is the footprint markings that appeared on the floors of convenience stores and supermarkets during the COVID-19 pandemic. People interpreted these markings as indicators of social distancing and voluntarily adjusted their behavior, standing on the marked spots while waiting in line. This behavior has now become firmly established. To discover such universally applicable measures, research across psychology, physiology, and sociology is essential. One of ITLAB’s strengths is the freedom to identify what knowledge is needed and pursue research accordingly.

One of the major challenges in this research is that true effectiveness can only be verified through experiments in real vehicles, which in practice often requires commercialization. Yet commercialization is difficult without confirming effectiveness, creating a classic dilemma. At present, we are validating hypotheses using driving simulators while preparing for evaluations in real vehicles.

Most human behavior is habitual. People do not consciously think about every action they take. When there is a door, our hand naturally reaches for the knob, and we open it without thinking about which direction to turn it. Acting consciously all the time is exhausting, and to live comfortably, people inevitably rely on habits. Safe driving is no exception—habit formation is essential. Ideally, safe behaviors would be ingrained in the body without requiring conscious awareness. To achieve this without causing annoyance, we may need to move beyond conveying information through audio or visual displays, and instead explore physical AI approaches in which machines directly interact with the human body to some extent, allowing conditions and actions to be learned unconsciously.

Since joining ITLAB, my goal of achieving zero traffic accidents hasn’t changed. ITLAB allows me to freely define research themes aligned with that goal, while also offering flexibility in working hours, locations, and conference participation. I believe this freedom is one of ITLAB’s greatest strengths.