German automotive giant BMW Group is teaming up with French AI specialist Mistral AI to revolutionise virtual crash testing. The aim is to speed up the development of safer vehicles by leveraging artificial intelligence capable of analysing vast datasets.
In the automotive world, safety is paramount. For BMW Group, virtual crash simulation is a cornerstone of its development process. Thousands of virtual simulations are conducted every week, generating an enormous amount of data – over a petabyte to date! It’s in this complex environment that artificial intelligence promises to make a significant difference.
AI: A New Tool for Automotive Safety
The partnership between the German manufacturer and Mistral AI, a specialist in artificial intelligence models, aims to enhance the quality, accuracy, and speed of crash simulations. This marks a significant first step in integrating AI into more specialised areas of automotive development, and more broadly, across BMW Group’s entire value chain. The idea is to create ‘tailor-made’ AI models capable of understanding and analysing the specific technical nuances of automotive engineering.
Industrial Data for Cutting-Edge AI
“For BMW Group, the use of industrial data is a key factor in translating artificial intelligence into value creation,” explains Dr. Franz Decker, CIO and Senior Vice President of the group. The alliance with Mistral AI combines BMW’s extensive engineering datasets with the French company’s model training capabilities. The expected outcome is specialised AI that can assist engineering teams with particularly demanding development tasks.
A Staggering Volume of Data
The complexity of crash simulations at BMW, with thousands of virtual tests conducted weekly, generates a phenomenal amount of technical data. This information, accumulated over years, offers valuable insights into the behaviour of vehicle structures and materials during an impact. This data treasure trove provides a unique foundation for training an industrial AI model capable of continuous learning and improvement.
‘Large Industry Models’: The Future of AI in Industry
To deploy this approach on a large scale, BMW is focusing on ‘Large Industry Models’ (LIMs). Unlike general-purpose AIs, these models are specifically trained on engineering and simulation data unique to the automotive industry, particularly from safety tests. LIMs thus incorporate highly specialised domain knowledge, requiring not only industrial data but also deep technical expertise and learning environments tailored to BMW’s development processes.
A Strategic Collaboration for Innovation
This partnership underscores the growing importance of industrial data for the next wave of value creation. It also strengthens BMW Group’s innovation and artificial intelligence ecosystem, positioning it at the forefront of technology in a sector where safety is critical. Integrating AI into crucial areas like crash simulation promises to accelerate the development of increasingly high-performing and, above all, safer vehicles for everyone.
Key Takeaways from this Partnership
- Accelerated Development: AI is expected to reduce the development timelines for new models by optimising crash simulations.
- Increased Accuracy: Fine-grained data analysis by specialised models should improve the reliability of simulation results.
- Enhanced Safety: The ultimate goal is to design even safer vehicles through a better understanding of impact dynamics.
- French Expertise: The collaboration highlights Mistral AI’s know-how in advanced artificial intelligence.
- Towards Industrial AI: This project paves the way for the use of AI models specific to other industrial sectors.
