JapanManufacturing2025 – ongoingHonda Motor
Generative AI and surrogate model for front-design changes that meet pedestrian-protection performance
Honda estimates its generative AI design tool cuts design development time by more than 40%
Problem
Front designs had to meet pedestrian-protection performance, and checking each design with CAE analysis could take about a day. Design changes went back and forth between designers and analysts.
Approach
Honda developed the system in March 2025 and applied it to actual vehicle development. A designer submits a 3D design and gets performance checks and revised proposals in a chat, using a large language model from OpenAI. A surrogate model stands in for CAE to judge performance almost instantly. Morphing, an image-processing technique, generates the revised shapes, because direct AI generation could not reach millimeter precision. Designers can set limits such as keeping the hood height.
Results
Takeaway
The surrogate model reached about 90% of CAE accuracy, so Honda still ran a final CAE check. The time saving is Honda's estimate, not a measured result.
Categories
- Size
- Large enterprise
- Industry
- Manufacturing
- Business function
- IT and R&D
- Innovation type
- Process innovation
- AI technology
- Generative AIPredictive analytics and ML
- Adoption
- Built in-house
- Evidence
- Company's own claim
Sources
Source summaries
The article, from the print edition of 2025-12-22, described Honda's chat-based generative AI system for front designs. It was developed in March 2025 and applied to vehicle development. It reported an estimated cut of more than 40% in design development time and surrogate-model accuracy of about 90% against CAE.
Full details are in the original sources.