Should AIs fear to be replaced by humans?

AI is a realm of paradoxes. On the one hand, a public space saturated with thundering prophecies and apocalyptic predictions announce either the golden age or the coming takeover by robots ready to destroy millions of jobs and supplant mankind. On the other hand, on the ground, not one but a wide variety of AIs are already there, in many activities and professions. They are already transforming our lives and ways of working. From the expert systems long used in the nuclear industry, for example, to the spam filter in our e-mail inboxes, not to mention electronic assistants and the countless systems that are already automating whole swathes of our daily lives.

AI in the automotive industry

The automotive industry is no exception. The epiphany of the autonomous car, the "level 5" Graal is constantly announced, and constantly postponed. Despite undeniable progress, who will ever want to entrust their lives to systems that will always lack the foresight to understand that a car must not stop because a passer-by is carrying a "stop" sign? At the same time, no one can ignore the spectacular progress and thousands of lives potentially saved by driver assistance systems. Who can deny the usefulness of automatic detection of cyclists to warn drivers that there is a risk of collision?

As much as in the product, it is on the design and production side that the AI revolution has begun for the automotive industry. AI is already at the heart of the Renault Group's industrial operations. It now pervades all its businesses, from design to logistics.

Where machine breakdowns used to cost millions of euros in losses, training algorithms using statistical learning open the way to new solutions. Based on weak signals and vibrations picked up by sensors, maintenance operations can be triggered even before a problem occurs.

In addition to productivity gains, reductions in CO2 emissions are achieved, by optimizing parts delivery circuits from suppliers to factories, as well as truck loading. Designers also are already making the most of generative AI, stimulating their creativity by exploiting the possibility of obtaining images of vehicles, in the most diverse environments.

Let's make no mistake: these tools don't just fall from the sky. They are built by people. Their qualities and shortcomings are the result of those who develop them and those who use them. 

Human – machine connection 

Faced with the current technological acceleration, of which ChatGPT is nothing but the most spectacular symptom, we are convinced that our first challenge is to invent the right relationship between AI and the men and women who use them. What better reminder thereof than the blossoming of Silicon Valley companies specializing in "prompts"? Only the art of feeding the machine with the right instructions can unleash its power, in the service of a human creativity.

The same is true for the men and women on our production lines. Operators are becoming part of the machine learning operations, interacting with data specialists, and developing an almost artisanal ability to grasp how AI systems work in order to react more effectively. Take the case of a defect detection system: it is now part of the operator’s job to attribute a system malfunction to modifications as subtle as a change in lighting or exposure in relation to the context in which the AI has been trained: alongside advances in automation, it is the ingenuity and intelligence of human situations that are called upon more than ever.

AI the Renault Group way: humanized technology

For automakers, man-machine is also a product challenge. At Renault Group, we are convinced that AI can be a lever for humanizing technology. Cars become their users’ companions, learning from them, evolving in contact with them. Thus, automobile is about to enter the era of personalized experience. This is also about harnessing a competitive edge: no one is better placed than European automakers to design an "intelligent" car that is as close as possible to the culturally marked expectations of European consumers.

Our ambition is to make the Renault Group the first AI native automobile manufacturer. There is no AI without data and the huge work to collect them and harmonize them. By starting several years ago to centralize our industrial data we have set essential milestones. We have ensured their interoperability and fluid circulation by means of a digital twin project which is unique in its kind within the entire industry. Today, over a billion data are collected every day from all our industrial sites and operations.

The second challenge addressed by Renault Group is about innovation management. We are faced with the rapidity of technological leaps likely to affect all sectors, and even every aspect of human activity. The next breakthrough triggered by AI can pop up anywhere. The field to be covered is immense and no single player will be able to have all the answers and seize all the opportunities alone. Our response therefore involves a horizontal and open approach, closely associating our teams with the best tech companies to co-develop and co-create, but also to have one foot already on the next train to leave. Collaborating with teams from Google and Qualcomm is already a daily reality for our engineers.

This collaborative approach also guides us when we take part in the Confiance.ai program, within a group of major French academic and industrial players. In this research program launched by the French government, we pool cutting-edge scientific and technological know-how to design and industrialize trustworthy AI-based critical systems.

To cut it short, this is no time for autopilot mode! The only thing AIs should fear is that humans will let them down. Rest assured: Renault Group is not short of ideas!