Anthropic reveals: AI poses a greater threat to white-collar jobs than expected

Le brief IA que les pros lisent chaque soir
Les 7 actus IA du jour, décryptées en 5 min. Gratuit.
Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.
Choisis ton rythme
Gratuit · Pas de spam · Désabonnement en 1 clic
A Study That Challenges Certainties
Anthropic, the company behind the AI Claude, has recently published a study that questions the prevailing assumptions about jobs threatened by artificial intelligence. Contrary to common predictions, it is not only low-skilled jobs that are at risk. The study employs an innovative analysis method, intersecting the theoretical capabilities of AI with its concrete applications. Researchers observed not only what Claude is technically capable of doing but also what users actually request from it in their daily tasks.
This approach radically alters our understanding of the labor market in the face of AI. Until now, the dominant belief was that repetitive and low-skilled jobs would be the first to be affected by automation. However, Anthropic's study reveals a more complex and concerning reality for skilled workers. Developers, salespeople, and certain managers are already significantly exposed to this new dynamic. The threat of AI is no longer a distant hypothesis but a reality that is infiltrating daily work processes.
Skilled Jobs on the Front Line
Anthropic's study stands out for its focus on the concrete use of AI. Instead of relying on abstract criteria, it examines the real interactions between users and AI. This method has allowed for the creation of a "professional risk map" that highlights the jobs already impacted by AI.
Developers are among the professions most solicited by artificial intelligence. Claude is already being used to generate code, correct errors, and explain complex functions. This massive usage places programming in a zone of high vulnerability. Similarly, sales jobs are affected, with tasks such as writing emails, preparing pitches, and analyzing prospects increasingly delegated to AI.
Management is also not exempt from this transformation. Reports, summaries, and other internal documents are often processed by AI, fundamentally altering the nature of work. While this does not mean an immediate elimination of jobs, it leads to a reconfiguration where humans must adapt to increased pace and productivity.
The Refuge of Physical Jobs
In contrast, the jobs best protected from AI are not necessarily the most prestigious. Anthropic emphasizes that professions rooted in the physical world, such as cooks, mechanics, or bartenders, are harder to replace. These jobs require precise gestures, physical presence, and direct sensory perception.
A cook adjusting a sauce, a mechanic listening to an engine, or a bartender adapting their service to the ambiance of a crowded room are examples of tasks that still elude AI's capabilities. These situations demand human interaction and immediate adaptation that AI cannot yet replicate.
Anthropic's study highlights a paradigm shift in the perception of threatened jobs. It is no longer a matter of distinguishing between repetitive and creative tasks, but rather between digital and physical professions. This new perspective explains why some white-collar jobs are more exposed than manual jobs.
An Evolving Boundary
The advice from the CEO of Palantir, suggesting a shift towards manual jobs like plumbing to escape the impact of AI, resonates particularly in this study. Anthropic validates this idea but with a more structural explanation: the issue lies in the distance between digital work and the real world.
However, one unknown persists and could quickly change the game. If robotics progresses significantly, the physical boundary currently protecting certain jobs could collapse. For now, the study primarily underscores a disturbing reality for digital workers: the most threatened jobs are not always those one might expect.
Brief IA — L'actualité IA en français
L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.