The idea of a car designed entirely by artificial intelligence isn't science fiction anymore. Car companies have started using AI tools to radically speed up their development processes, and the results could change the cars we drive forever. But the shift also carries significant consequences for the people who design and build them today.

According to a recent episode of The Vergecast, contributor Tim Stevens walked through how automakers are incorporating AI into everything from initial sketches to final production designs. The conversation covered the current state of automotive design, why clay models are still relevant, and the risks to the jobs pipeline in the industry.

How cars get designed today

Car design has historically been a slow, labor-intensive process. Designers sketch ideas, sculpt clay models at 1:4 scale and then full size, and iterate through countless physical prototypes before a single production vehicle rolls off the line. That workflow can take years.

AI tools are now being used to compress that timeline. Generative design algorithms can produce hundreds of styling alternatives in the time it used to take to draw one. Engineers can feed in constraints like weight, aerodynamics, and material properties, and the software returns optimized shapes that humans might never have conceived.

Stevens noted on the show that the technology isn't just about making pretty pictures. It's about exploring a much wider design space much faster. A human designer might have time to try a dozen variations of a front grille. An AI can generate thousands and rank them by performance criteria.

Clay models still matter

Despite the push toward digital tools and AI-driven design, clay modeling remains a critical step. There is something about seeing a full-scale physical model under real lighting that no screen can replicate. Automakers still rely on clay sculptors to evaluate proportions and surfacing before committing to tooling.

But the role of those sculptors is changing. Instead of starting from scratch, they now often work from 3D data generated by AI systems, refining and validating rather than inventing. The pipeline is becoming a hybrid human-machine workflow.

Jobs pipeline risk

The shift has obvious implications for employment. If AI can generate many designs that used to require teams of junior designers, fewer entry-level positions may be available. The pipeline for training the next generation of automotive designers could narrow. Stevens pointed out that the industry is aware of this risk but hasn't solved it.

At the same time, the technology could create new roles. Automakers will need people who can train and fine-tune AI models, interpret their outputs, and make subjective judgments about brand identity and aesthetics. The question is whether those new jobs will outnumber the ones displaced.

Software-defined cars change the equation

The adoption of AI in design runs parallel to another major shift: the move toward software-defined vehicles. Modern cars are increasingly platforms for software, with features added and updated over the air long after the vehicle leaves the factory. That changes how designers think about hardware, because the physical design must accommodate future software capabilities.

AI tools can help model those interactions early in the design process, simulating how a particular camera placement or sensor array will affect an autonomous driving system months before a physical prototype exists. The integration of hardware and software design is becoming a single problem rather than two sequential ones.

Regulation and safety

Safety regulations remain a major constraint. Every vehicle sold must pass crash tests, pedestrian protection standards, and a host of other requirements that vary by market. AI design tools can help engineers meet those rules more efficiently by optimizing structures for impact performance while minimizing weight. But regulators are still catching up to the idea that a computer generated part of the design. The episode discussed whether existing certification frameworks are adequate for AI-designed components.

Will AI homogenize design?

One concern that came up on The Vergecast is whether widespread use of AI will lead to cars that all look the same. If every automaker feeds similar constraints into the same kind of generative tools, the outputs might converge on similar shapes. Brand differentiation could suffer.

Stevens noted that the answer depends on how companies constrain their AI systems. If a brand gives the AI a strong identity brief — "look like a BMW" — the results will stay within a recognizable family. If the AI is allowed to optimize purely for aerodynamics and cost, the results may be harder to tell apart. The responsibility falls on human designers to set the guardrails.

A broader AI news roundup

The episode also covered several other AI developments that paint a picture of the industry's current mood. Verge reporter Hayden Field discussed the competition between Claude Code and Codex, two AI coding assistants, and the growing tension between Anthropic and the US government over Pentagon contracts. Anthropic has said it will not work on certain military AI applications, while the government continues to push for access.

Field also gave a vibe check on OpenAI, describing the atmosphere inside the company as shifting between optimism and internal turmoil. The PR battle between AI boosterism and doomerism continues, with no clear resolution.

A listener hotline question asked whether the recent wave of tech layoffs is really about AI, or if AI is just a convenient explanation. The panel discussed the evidence, noting that many companies have used the AI narrative to justify restructuring that was probably going to happen anyway.

What comes next

The automotive industry's experiment with AI design is still in its early stages. Few production cars on the road today were designed entirely by AI, but the tools are being used in earnest. The next few years will show whether the technology delivers on its promise of faster, cheaper, better vehicles, or whether the loss of human craft proves too great.

For now, the message from The Vergecast is clear: AI is not replacing car designers tomorrow. But it is already changing how they work, and the cars that reach the road in five years will look different because of it.