Adoption of artificial intelligence in business, how to use it

Traffic is accelerating. Since the release of ChatGPT by OpenAI a year and a half ago, genetic artificial intelligence (AI), capable of imitating human logic, has become widespread among the general public. She also made her way into the business. After the first tests, it's time to get down to business. And the stakes are high, because the global technological race continues to intensify. “We should prepare French and European companies for major geopolitical and investment battles”, announces Rédouane Bellefqih, CEO of Deloitte Consulting France. But to be able to ride the wave of AI, the issue “It cannot be approached exclusively from the perspective of innovation, but in a broader way, to embrace the whole company and its place in society”he adds.

And you can also start by asking yourself basic questions, even before developing an AI project. “The goal is to create value, with a certain number of safeguards to be put in place”, recalls Béatrice Kosowski, president of IBM France. And therefore find out how to use this technology to “increase” the individual in his role within the organization. First step for a business: “Think about the alignment, in relation to its strategy and top priorities, for what it can achieve with AI”, notes the director of the digital giant. Then gather the occupations to develop a uses map. Some companies have already made good progress on this path, others are just beginning to explore it…

Variety of models

More detail, “When a company wants to develop an AI system, it also faces strategic choices, such as whether to choose the closed, proprietary model or the open model.”adds Laurent Daudet, co-founder of LightOn, a French gem that develops language models and helps companies personalize them. “We also have a very open multi-model approach, but we campaign for small, targeted models that have the advantage of frugality and lower cost”adds the IBM France manager, who has partnered with French start-up Mistral AI, among others.

“What will be decisive for companies, but also for start-ups [qui développent les solutions d’IA]it's about scalability and speed of execution.”, decides Rédouane Bellefqih. But beyond the technical functional aspects, it is the whole “strategic human dimension, which deeply affects the company” which must be taken head on by leaders, he insists. Especially from “AI is no longer a matter of IT departments but of professions”from R&D to production, including marketing and legal issues, notes Laurent Daudet.

Train and expand pools

The whole challenge, in fact, to benefit from the AI ​​revolution, lies in skills. But these are missing. According to a recent report by the Interministerial Committee on Genetic Artificial Intelligence, 80,000 jobs are needed for its development and growth. Hence the need to educate the general population, specialists and those whose professions will lead them to use it.

For example, at Orange, about 2,000 people, mostly engineers, historically used AI tools. “The breakthrough with genetic artificial intelligence is the much easier access of artificial intelligence to a very large number of professions, whose activities will be modified.”, reports Vincent Lecerf, HR director at Orange. This is why the team is working on creating awareness among employees on different applications. “We have trained 27,000 people”, so confirms this HR manager. Similarly, Orange has created its own training center for data scientists. “There is a tension in the market, between an exponential need and a human resource that is not growing at the same speed. So we are forced to develop our own upgrades”sums up.

For its part, the University of Paris-Saclay aims to “in a few years whoever graduates [de Paris-Saclay] trained in artificial intelligence “, says Sarah Cohen Boulakia, a professor at Paris-Saclay University. If future AI experts are necessary, the challenge is also to train AI that is oriented towards professions – those of lawyers, doctors, biologists, chemists, etc. “These are artificial intelligences that will adapt to specific areas where soft skills will play a role in particular”adds the deputy director of the DATAIA Institute responsible for AI training.

In all cases, to bridge the gap between needs and available skills, it is necessary to expand the teams to include new profiles. This issue is also moral. Because if genetic artificial intelligence is now in everyone's hands, it is still built from the same profiles – those of engineers, and often men, warns Marie Even, deputy general manager of Cdiscount. “AI models are only designed by half of humanity, obviously with biases. The lack of women in this field is alarming”she argues.

Watch the round table in the video below Embedding artificial intelligence in business, the big challenge of 2024 »

Leave a Reply

Your email address will not be published. Required fields are marked *