AI does not destroy jobs but reveals the obsolescence of recruitment still based on CVs, at the risk of transforming a promise of opening the labor market into a new form of selection.
“Artificial intelligence is destroying jobs”, this is an alarmist speech in keeping with the times. However, this is a legitimate concern, at the height of the current revolution but which focuses on the wrong part of the problem. AI does not primarily impact the number of positions; it intervenes in a more discreet area, but much more decisive: access to opportunities. What she redefines is not just the work but the front door that leads to it.
The CV, this prism that has become too narrow
For decades, talent selection was based on an almost automatic gesture, that of looking at a CV like reading an identity card. We tracked down coherence, good names, agreed paths. The diploma served as a moral guarantee, linearity as proof of seriousness. This system was not perfect, but it offered a benchmark in a slower world, where experience accumulated over the years.
Today, this mechanism is cracking everywhere. Unemployment reached 7.7% in France. According to APEC, executive employment fell by 8% in 2024 and is expected to decline by a further 4% in 2025 due to economic and political uncertainties. Young graduates and those over 50 are at the forefront of this trend. The climate remains gloomy and many recruitments are currently frozen.
The real labor market is no longer national. It’s not even continental anymore. It is global, distributed, fragmented, fluid and in the face of it, our hiring reflexes remain abnormally fixed.
AI, a new contribution accelerator
This reversal would already be considerable in itself. But AI adds a second upheaval, even more radical because it opens up qualified professions to candidates who until now had only theoretical access.
With the right tools, a self-taught person can analyze complex data, write robust technical documentation, and code a functional prototype. A single parent living far from urban centers can collaborate with an international team, producing work once reserved for prestigious graduates. The gap between assumed skills and actual capabilities is narrowing. The CV loses its status as objective truth because what counts now is what we know how to do today, not what we did yesterday.
Recruitment systems still trapped in the past
But if AI opens doors, recruitment algorithms sometimes close them immediately. Numerous automated filters immediately eliminate atypical career paths, late retraining, and conflicting careers. They learn from past decisions and repeat them, sometimes with cold consistency. In a world where skills are in flux, these tools continue to favor the same inherited signals, the same social codes, the same obsolete certainties.
This is where the major risk lies; not that of massive unemployment orchestrated by machines, but that of silent segregation, where only those who already had the codes take advantage of new opportunities. AI could expand the job market. She could also close it.
Fairness does not spread by itself
Contrary to what we would like to believe, AI is not a magical equalizer. It amplifies what we give it and if we want it to really broaden access to qualified work, we must review the way we evaluate candidates. Test the resolution of a problem rather than fluency in an interview, probe reasoning rather than gauge a posture. Allow the use of AI during tests, because this is how modern work is built.
Some companies that operate asynchronously have understood that the clarity of written reasoning, the ability to organize alone, to learn quickly, to collaborate remotely are worth a thousand times more than a diploma displayed in the margins. Competence becomes alive again, measurable in action, tangible in the way of doing things.
A societal issue, not a simple HR optimization
This transformation is not anecdotal. It touches on a profound principle: who do we give a chance? Distributed work and the search for very specific skills create an unprecedented window of opportunity. For the first time, entire populations from rural areas, developing countries, or far from academic networks can connect to a global job market.
It is still necessary for hiring systems to stop functioning as social customs but also to identify the causes so that these principles are transcribed from an operational point of view. It is a living process where catapulting new tools, however powerful they may be, requires integration commensurate with the situation: almost a third of French HR managers describe a bad experience with the technological tools available, and poor integration of the latter, again according to our data. Training and thinking about your technological contribution is an issue just as crucial as innovation itself.
A choice of civilization
Basically, AI asks us a simple question: do we want to expand the circle of possibilities or protect it? Do we want to evaluate individuals on their potential or on their conformity?
Technology will not decide for us. She doesn’t say who is worth hiring. It only forces us to recognize that yesterday’s criteria are no longer appropriate, and that the emerging labor market will be profoundly unequal if we persist in recruiting as if AI did not exist.
