TOKYO - The development of autonomous driving technology is entering a new phase, driven by advances in generative AI. A white vehicle seen driving through central Tokyo without a driver was recently unveiled as Japan's first public demonstration by Waymo, a subsidiary of Alphabet.
The event marked a turning point, as startups and major players alike accelerate efforts toward what is being called "Autonomous Driving 2.0"—a new stage powered by AI.
Waymo, which began its self-driving taxi service in Arizona, is now expanding to San Francisco, Silicon Valley, and Los Angeles. In San Francisco, the service is gaining popularity even among visiting Japanese businesspeople, becoming something of an attraction.
On May 3rd, Waymo and Toyota Motor announced a basic agreement to jointly develop autonomous driving technologies. This marks a shift in attitude by Toyota, which had previously been wary of Google's data-collecting capabilities. The move suggests that Toyota is taking autonomous driving more seriously.
Interest in this field is also growing among AI researchers. Nvidia CEO Jensen Huang, a prominent figure in AI development, has repeatedly stated that autonomous driving is the next major frontier after generative AI.
The term "Autonomous Driving 2.0" reflects growing recognition that the previous model, "1.0," had reached its limits. Although research began more than a decade ago—Google’s efforts date back to 2013—the technology remains limited to specific urban areas, with little progress in making it widely available in consumer vehicles. Despite investments exceeding 100 billion dollars (about 14 trillion yen), widespread adoption has not yet materialized. One reason is that training systems using real-world driving data struggles to handle rare or unexpected scenarios.
In contrast, Autonomous Driving 2.0 introduces a major structural shift. The older 1.0 systems separated recognition, decision-making, and control into distinct AI modules. The new model consolidates all of these functions into a single large-scale AI system. This simplification not only makes development and updates easier but also enhances performance.
The concept is similar to how ChatGPT integrates various data inputs—text, images, and audio—to draw conclusions. That same architecture is now being applied to vehicles. The emergence of generative AI has created favorable tailwinds for this approach.
Among the key players leading this shift is Tesla. The company recently announced that it would procure over 3 billion dollars’ worth of advanced semiconductors from Nvidia, which serve as the brain of these next-generation systems.
Source: テレ東BIZ
