Future Says S4E2: Transforming Renault into a Tech Company with Luc Julia
Over the past two years, Luc Julia has been applying a wealth of Silicon Valley experience to his role as chief scientific officer at Renault, the French carmaker that sold over two million vehicles in 2022. Reporting directly to the CEO, Julia is at the heart of a mission that aims to transform Renault from a car company working with tech to the opposite: a tech company working with cars. This ambitious strategy is a direct response to the dramatic disruption being felt throughout the automotive industry.
On the latest episode of Altair’s Future Says podcast, it’s no surprise to find that Julia is a passionate advocate of technology as a driver for change. However, he’s equally keen to highlight the dangers of overhyping both the possibilities and pitfalls of the latest generation of accessible artificial intelligence (AI).
Harnessing the Spirit of Silicon Valley
During his three decades in California, Julia worked for enterprises ranging from startups to tech giants. He was the chief technology officer and senior vice president for innovation at Samsung, chief technologist at Hewlett-Packard, and ran the Siri project at Apple. Along the way, he also found the time to write two books, including one titled “There Is No Such Thing as Artificial Intelligence.”
Julia starts the interview by explaining what drew him to the specially created C-suite role at Renault. Above all else, he was keen to apply his Silicon Valley experience in his homeland of France. And in a company the size of Renault, he sees no shortage of opportunities to apply innovative technologies "not only in the cars themselves, but also in the conception of the car, in the building of the cars, and in any support function.”
To illustrate further, he reels off potential applications. In the design phase, these include using generative AI to improve the creative process. In production, creating digital twins of factories to inform predictive maintenance programs. In the car itself, gait recognition could be introduced, enabling a driver to unlock and personalize a vehicle via their body shape and walking style.
Autonomous Driving and Road Safety
The obvious starting point for deploying AI is to identify the repetitive processes that can be automated, and where that automation will provide most help to the people currently doing those tasks. That will give them the freedom to concentrate on jobs where AI can’t match peoples’ unique skills and intuition.
In fact, Julia is keen to stress that technology has its limits. Fully autonomous driving is a good example of these limitations. In contrast to some key players in the industry, he doesn’t expect Level 5 (fully autonomous driving) to reach the mass market. However, he’s confident that Level 4 is realistic, and will ultimately deliver huge benefits for road safety.
The challenges surrounding Level 5 aren’t just technical – affordability is important too. “I’m always doing technology for real people,” Julia said. Even if Level 5 remains out of reach, the rollout of more autonomous driving features will present exciting opportunities to reimagine the in-car experience as an extension of the home. Before long, Julia speculates, the car’s windows might also become a medium for augmented reality, replacing or enriching the occupants’ view of the outside world.
The Present and Future of AI in the Automotive Industry
His enthusiasm for the possibilities of AI is tempered with a dose of realism. AI is indeed an incredibly powerful tool or, more accurately, toolbox. For those that fear the consequences for humanity, the bad news is that AI can outperform humans in many tasks. But Julia compares AI to a hammer. Obviously a hammer is designed to outperform a human hand; people design tools because they help us do things better and more efficiently. The same is true of AI – it’s designed to do things better. And just like a hammer, AI can be used for good or ill. The responsibility lies with the person using it, not the tool itself.
That’s why we need to educate ourselves, Julia says. We must be able to distinguish between the real AI and what Julia refers to as “Hollywood AI.” That also means keeping AI in perspective. As a technology that dates back to 1956, it can hardly be described as revolutionary. Even the latest wave of generative AI is not particularly “new,” since it’s built on decades upon decades of research and testing. The real change lies in the accessibility of tools such as ChatGPT. “If there is a revolution, the revolution is in the prompt.”
In common with other Future Says guests, Julia doesn’t love the term “artificial intelligence.” Instead, he’s much happier with “generative AI” since he believes it better describes how the technology works and what it delivers. That’s because, for all its capabilities, AI isn’t creative in the way that the human mind is. ChatGPT generates output based on the data it is fed. It is largely mimicry, even if that mimicry seems indistinguishable from human-made work.
Julia recognizes that the growing accessibility of AI increases the potential for users to do harm. But he’s equally convinced that we shouldn’t be rushing to introduce new regulations based on misunderstandings and misguided fears of an AI-led takeover of humanity. The issue here isn’t with regulation per se, but with the speed and direction of initiatives such as Europe’s forthcoming AI Act. He believes it would be far better to wait until we have a more complete understanding of what the technology can do. Moreover, regulation should be targeted at undesirable applications, not the technology itself.
Conclusion
Julia thinks we are probably at the peak of the current AI hype cycle. We’re beginning to recognize the limitations of heavyweight tools such as ChatGPT. Essentially, ChatGPT draws on the entire internet as its source material. That makes it too cumbersome for many applications. We also need to consider the environmental costs involved. As the size of the data increases, so does the amount of power needed to train it. Sustainability must be part of the calculation.
As a result, Julia expects to see smaller, more focused, and more specialized AI applications emerging. To stretch the hammer metaphor a little further, we’ll see generative AI adding further tools to the toolbox. At Renault, that should help the company to flourish in an industry where virtually every accepted truth about designing, building, and selling cars is being challenged. Meanwhile, in the wider world, getting the best out of AI is a question we all need to face. In clear and convincing fashion, Julia has no doubt where the future of AI lies. “We decide. We are in charge. I have the handle of the hammer in my hand.”
Click here to listen to the full episode with Renault’s Luc Julia. To learn more about the Future Says series and browse previous episodes, visit https://altair.com/future-says.
Future Says season four is proudly sponsored by Oracle. Oracle offers integrated suites of applications plus secure, autonomous infrastructure in the Oracle Cloud. For more information about Oracle (NYSE: ORCL), please visit www.oracle.com.
