GM and Nissan: AI Redefines Car Design

AI Revolutionizes Automotive Design

In a global context marked by trade wars and uncertain demand, automakers are increasingly relying on artificial intelligence to reduce development timelines. The world of automotive design is filled with advanced 3D visualization tools and virtual sculpting platforms, but most new cars still begin their existence as sketches. These sketches traditionally undergo countless iterations and refinements before being transformed into 3D models manually, with some remaining in the digital realm while others are sculpted in clay to better visualize lines and profiles. This design and development process often takes five years or more.

This means that many new cars arriving at dealerships this summer were first sketched in 2020 or 2021, at a time when incentives for alternative fuels were widespread, electric vehicle charging stations were multiplying, and the days of internal combustion engines were numbered.

Impact of U.S. Policies

Today, everything has changed. The second term of the Trump administration put a brake on all kinds of incentives for electric vehicles while imposing tariffs and import/export restrictions. Automakers that had promised to transition entirely to electric by the end of the decade now find themselves integrating engines into everything that moves, and factories are quickly being repurposed to avoid the worst of the import restrictions.

In this context, we are witnessing a boom in agentic AI, which more and more manufacturers are leveraging to reduce this 60-month design and development window. As with most aspects of AI, the potential is immense, but some consequences are also concerning.

AI at General Motors

At GM, the process of developing new cars benefits from an injection of AI during the design phase. Dan Shapiro, a creative designer at General Motors, explained the workflow, which always starts with human design. “That’s what sketches are for,” he said, “and AI helps us visualize it earlier.”

By feeding hand-drawn sketches into a commercial tool called Vizcom, Shapiro was able to create a fully realized 3D model and animation in a matter of hours, a process that he said previously took “multiple teams several months.”

Shapiro's example was a concept car with aggressive lines that would fit right in on the streets of Night City. By writing prompts such as: “Create a dynamic action view of this Chevy concept vehicle... Empty elevated streets. Modern city,” he produced a simple animation. Soon, it was driving on perpetually wet roads, typical of a cyberpunk future.

For now, these animations are used only internally as rolling mood boards to help GM teams visualize what works. Shapiro emphasized that it is still human designers who shape things, not AI: “We are still the monks who decide what looks like a Buick, a GMC, a Cadillac, and in this case, a Chevy.”

But AI is also having an influence in this area.

Neural Concept and CFD

Computational fluid dynamics (CFD) is the science that determines how a fluid flows around a given shape. CFD helps electric vehicles travel a bit further on a charge, and large trucks offer slightly improved wind resistance. Since 2018, a Swiss company called Neural Concept has introduced the power of neural networks into the art of CFD. Tasks that previously took hours on supercomputers can now be simulated in minutes on GPUs like those from Nvidia.

Neural Concept has applied its technology to everything from family sedans to Formula 1 cars (Williams Racing is a client), and while most of its clients prefer to remain anonymous to keep the details of their design tools and processes secret, Jaguar Land Rover (JLR) recently praised this technology. At Nvidia's GTC this year, Chris Johnston, senior technical specialist at JLR, stated that aerodynamic tasks that previously took 4 hours are now completed in 1 minute.

GM is following the same path, developing what it calls an “AI-powered virtual wind tunnel.” Scott Parrish, technical researcher and lab manager at GM R&D, showed me a demonstration. “We have developed an AI model to provide near-instantaneous drag predictions,” he explained. Designers and engineers can adjust surfaces and receive feedback almost instantly.

It’s not just cars that are being redesigned. GM's process is evolving as well. While designers previously handed models off to CFD engineers, who would test for days or weeks before providing feedback, the process is now more iterative. Additionally, since designers can quickly produce 3D models, CFD work can begin earlier.

However, these automated procedures are not perfect. “We are building autonomous systems that design cars with strong human supervision,” said Pierre Baqué, CEO and co-founder of Neural Concept. “The value lies in the combination of AI speed and human judgment, not in removing the human from the equation.”

The Importance of Code

The appearance of a car and its ability to cut through the air are not the only aspects contributing to a five-year development roadmap. Coding is becoming an increasingly important task. The push for software-defined vehicles means more complex integration efforts that have delayed launches and cost billions. AI is also seen as a potential asset in this area.

At Nissan, the focus is on automating some repetitive software development tasks, such as unit testing. Takashi Yoshizawa, an executive at Nissan in charge of software-defined vehicles, explained to me that these code generation tools “improve both development speed and quality.”

Streamlining Workforce?

A common refrain among companies venturing into AI is that they will enhance worker productivity by eliminating repetitive tasks, not by reducing headcount. GM representatives have been adamant about this. “This touches on a concern for many people, but the way we are actually leveraging AI allows people to do what they really came to GM for,” said Bryan Styles, director of design innovation and technology operations at GM Global Design.

Pierre Baqué of Neural Concept expressed similar sentiments for his clients: “Our platform is designed to amplify engineering teams, not to reduce them.”

Matteo Licata is not as sure. A former automotive designer, he is currently a professor at IAAD (Istituto di Arte Applicata e Design) in Turin. “Jobs in design studios may not disappear immediately, but in my opinion, only a fool would believe that such productivity gains won’t affect the number of employees in a studio in one way or another,” he said.

This has more troubling implications for Licata's students. “Getting into automotive design was already very difficult before AI, and now it’s only going to become more complicated,” he added.

Agentic Agility

Whether AI is an asset or a disaster largely depends on manufacturers' caution in its deployment. Some show better judgment than others. Dodge recently released “old family photos” of its most popular models from 20 years ago. In reality, the AI-generated images barely resembled reality.

Despite these marketing missteps, the current goal is speed. AI injections into GM's design process are already being used for its next-generation cars, but no one there is willing to comment on their market launch date. For its part, Nissan is aiming for a 30-month target for its new cars as it strives to regain momentum in the U.S. market.