Brief IA

Anthropic Refines AI Risk by Profession, 30% Less Exposed

🤖 Models & LLM·Tom Levy·

Anthropic Refines AI Risk by Profession, 30% Less Exposed

Anthropic Refines AI Risk by Profession, 30% Less Exposed
Key Takeaways
1Anthropic publishes a method that combines theoretical AI capabilities and real-world applications to assess the exposure of jobs to automation
230% of workers are in jobs identified as low exposure, such as cooks, motorcycle mechanics, lifeguards, and bartenders
3Computer programmers and customer service representatives are among the most exposed professions
4The list of the 10 most exposed jobs is available, and the analysis is expected to evolve
💡Why it mattersThis approach nuances the expected impact of AI on employment and allows for anticipating the most affected sectors.
Le brief IA que lisent les pros

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

📄
Full Analysis

Anthropic details a method that combines theory and observed uses of its products to estimate the exposure of professions to automation. The company anticipates more gradual effects than shocks like COVID and immediately identifies a large share of jobs that are less affected.

Anthropic anticipates a gradual evolution of jobs in the face of AI

Anthropic believes that the effects of AI usage on employment will not be as abrupt as those related to COVID, and may instead resemble the transformations brought about by the Internet or trade with China. At the same time, the company identifies jobs with low exposure, representing 30% of workers. It explains this lower exposure by the rarity of these activities in the data used. Examples cited include cooks, motorcycle mechanics, lifeguards, and bartenders.

A measure based on usage and application limits

Anthropic emphasizes that there may be gaps between what AI could theoretically do and what is actually implemented. The company highlights several barriers: model limitations that prevent the exploitation of certain capabilities, legal constraints, specific software requirements, and human verification steps that may slow down automation. Its new method thus aggregates the theoretical capacity to automate tasks and the actual usage data of its products, derived from the analysis of interactions with Claude and its API.

Anthropic publishes its method and identifies the most at-risk professions

The details of this methodology were published in March. By applying it, Anthropic has compiled a list of the professions most exposed to automation, which notably includes computer programmers and customer service representatives at the top. The list of the 10 most exposed jobs is available. Anthropic indicates that it aims to identify potential future economic disruptions in advance and clarifies that its analysis is expected to evolve. This initiative comes as artificial intelligence spreads across many sectors and raises concerns about employment.

Brief IA — L'actualité IA en français

L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.