A new study by the Bank of France reveals that automation does not lead to a noticeable overall decrease in the number of employees in companies. In other words, the total number of workers remains stable. However, organizations that use AI heavily are reducing their labor needs by limiting new hires and not replacing employees who leave. This uneven effect disproportionately affects younger workers. Between November 2022 and March 2026, permanent jobs for people aged 21 to 29 in the most AI-exposed sectors experienced a 28 percent delay in being filled. This gap shrinks for those in their 30s and disappears completely for those over 50. The technology sector, which has been at the forefront of AI development, is seeing the most significant declines. Large language models are taking over basic tasks such as writing simple code and creating documentation, which were traditionally done by entry-level workers. In France, employment for computer scientists under 30 has dropped by 7.4 percent in one year, according to recent data from INSEE. Similar trends are observed in the United States. The Stanford Digital Economy Lab estimates that hiring for software developers aged 22 to 25 has fallen by 20 percent since the end of 2022. These figures suggest that AI in IT departments is replacing the skills learned by junior workers during their education, but it cannot replicate the expertise gained through years of experience by senior staff. As AI takes over routine tasks, the way skills are developed within companies is also changing. Consulting firms and ESNs (external service providers) expect a shift in junior roles. Without the need to collect and format data, younger employees are now taking on tasks such as verification, strategy development, and customer interaction much earlier in their careers. "AI is a technology that will naturally allow us to increase our productivity, that is, to do what we did yesterday but especially to do things differently," said Mehdi Houas, President of Numeum, during recent meetings of the Medef, a major French business organization. However, this shift in productivity also requires changes in how managers operate. Industry experts suggest several strategies to adapt to this change: focusing on developing reflective intelligence rather than testing memory, using an inverted teaching approach where managers help teams think critically about AI use, and updating certification standards to emphasize mastery of large language models (LLMs). The Director General of the Treasury warns of a potential loss of experience among new generations of workers. IT departments now face the challenge of training future experts without relying on the traditional phase of learning through basic coding tasks. This shift requires new educational approaches and training programs to ensure that upcoming professionals can develop the necessary skills without the usual hands-on experience with fundamental coding.