The debate is no longer theoretical. According to the latest McKinsey Global Institute study published in March 2026, approximately 30% of current working hours in advanced economies could be automated by generative artificial intelligence by 2030. This figure, up from 2023 estimates, reflects the lightning-fast acceleration of language, vision, and reasoning models' capabilities. The most exposed sectors are no longer just manual repetitive jobs — as previous studies predicted — but also intellectual professions with a strong editorial, analytical, or administrative component: accountants, corporate lawyers, financial analysts, technical writers, translators, call center agents, and a significant portion of support functions in human resources and marketing.
The numbers are staggering. Goldman Sachs estimates that 300 million full-time jobs globally are "exposed" to generative AI. The International Labour Organization (ILO) published a report in February 2026 stating that 14% of jobs in OECD countries have a high risk of complete automation, while an additional 32% will undergo profound transformations. In France, Dares figures 4.2 million positions where at least half of the tasks could be performed by AI within three years. Accounting, data entry, legal secretarial, standard translation, and telemarketing jobs are among the professions most threatened by drastic workforce reductions.
But there is a bright side to the coin. AI doesn't just eliminate jobs: it creates radically new ones. The profession of "prompt engineer" — non-existent before 2022 — now boasts average salaries of 95,000 euros gross annually in Europe. "AI trainers," responsible for refining and evaluating models, have become a profession in their own right, with over 200,000 positions created worldwide in 2025. "AI ethics officers," "machine learning operations engineers" (MLOps), specialists in "AI-augmented design," and AI transformation consultants are among the most sought-after profiles in the market. LinkedIn reports that job postings mentioning AI increased by 340% between 2023 and 2026.
Reskilling strategies are being organized. In France, the Plan France 2030 has allocated 2.5 billion euros for training in digital and AI skills. Universities are multiplying hybrid programs: HEC is launching an MSc in "AI & Business Transformation," École 42 offers free prompt engineering modules, and Pôle Emploi (now France Travail) trained 150,000 job seekers in generative AI tools in 2025. The consensus among experts converges: the jobs that will resist best are those combining creativity, empathy, ethical judgment, and coordination skills — abilities that AI has not yet mastered. Artisans, caregivers, teachers, social workers, and field workers retain a structural advantage.
The real challenge, as highlighted by the World Economic Forum's 2026 Davos report, is not to fight AI but to integrate it. Workers who master AI as a productivity tool — and not as a substitute — see their market value explode. A developer assisted by Copilot produces twice as fast. A financial analyst using reasoning models processes ten times more data. A creative who effectively prompts Midjourney or DALL-E delivers proposals in hours instead of days. AI is not the end of work: it is the end of work as we knew it. And for those who can adapt, it is the beginning of an era of unprecedented opportunities.
