Japanese carmakers introduce generative AI to vehicle design

Chiang, Jen-Chieh, Taipei; Peng Chen, DIGITIMES Asia 0

Credit: Honda

Many automakers have been exploring how they can develop vehicle models with the assistance of generative AI. Japan-based Honda, Sony Honda Mobility (SHM), and Subaru have adopted the technology.

According to Nikkei, Honda revealed the Sustaina-C Concept, a small-sized EV, at the Japan Mobility Show this year for the first time. The company said designers have started to use image-generative AI to help them form ideas while not for official proposals.

The designer of Sustaina-C Concept said generative AI has evolved rapidly since the second half of 2022. The technology seems to have reached the singularity.

At the mobility show, Honda invited attendees to use Stable Diffusion, an image-generative AI tool developed by startup Stability AI, to design vehicles on their smartphones. Although the process was different from creating a real car, it gave consumers an idea about how generative AI can reduce the design time.

Toyota is also trying to leverage image-generative AI tools to design its cars. Subaru limits the adoption to specific departments to avoid infringement caused by the content that AI creates.

According to Japan-based Nikkan Jidosha Shimbun, Sony Honda Mobility, a joint venture between Honda and Sony, plans to use the large language model (LLM) of generative AI to develop the autonomous driving system and advanced driver assistance system (ADAS) for its EV Afeela.

Sony Honda Mobility aims to start selling Afeela in 2025 and begin deliveries in North America in spring 2026. The car's ADAS and autonomous driving systems will gather a large amount of information about the vehicle through sensors, cameras, and telecommunication systems. The data must be processed immediately to predict the movements of cars and pedestrians nearby.

Adopting LLM in machine learning for autonomous driving is expected to improve software development efficiency. LLM can help Sony Honda Mobility, which is late to the battery EV market, accelerate development, reduce the time to reach the break-even point, and set a foothold.