Scénario
Edition of October 30, 2026 · No. 99
FR
Photo d'illustration — IA et emploi : transformer ou détruire le travail ?
Scénario

Friday, science

AI and Employment: Transforming or Destroying Jobs?

Will AI cause net job losses or primarily transform the nature of work over the next 2 to 5 years?

Publié le 30 octobre 2026

The point, in plain terms

The emergence of autonomous AI agents threatens to automate millions of office tasks.

New artificial intelligence tools no longer simply answer questions. They now execute sequences of complex tasks completely autonomously. Nearly 5 million workers in France are directly affected in their jobs. The issue is not the end of work, but how quickly workers and companies can adapt to prevent widespread job losses.

Over the past two years, artificial intelligence tools have spread into daily life at a record pace. In France, nearly one in two residents now uses generative AI*—software capable of drafting text or generating images on demand. This adoption is outpacing that of smartphones in their day. While usage remains primarily personal for 42% of French people, it is also entering the workplace. Nearly 85% of 18-to-24-year-olds already use it regularly. This technological revolution is upending traditional workplace norms and raising an urgent question: will machines replace employees or simply reshape their day-to-day tasks?

The public debate often pits two starkly opposed views against each other. On one side, some predict a massive wave of layoffs in office jobs. On the other, commentators foresee mere time savings that will make workers more efficient without eliminating positions. In reality, the current situation is more nuanced. According to the Trésor-Éco* report No. 391 published in June 2026, no significant aggregate impact on overall employment is measurable yet. The French economy continues to create and shed jobs according to its usual cycles. For now, the technology’s overall impact remains invisible in national employment statistics.

Déjà abordé sur Scénario IA : risque d'extinction ou peur exagérée ? (13 octobre)

Yet a new generation of tools is entering companies and shifting the technical landscape: agentic AI*. Unlike a standard chatbot that waits for a prompt to reply, an autonomous agent can plan multiple actions on its own. It can debug code, book tickets, or sort files without continuous human oversight. According to the French Senate information report entitled “L’entreprise 5.0” published in April 2026, 16.3% of French jobs—representing nearly 5 million people—are exposed to this technology over a two- to five-year horizon.

Understanding it

Agentic AI operates like an assistant taking initiative.

Standard generative AI resembles an interactive dictionary: it responds when prompted. Agentic AI is more like an autonomous intern. Given a general goal, such as organizing a schedule or verifying invoices, the software breaks down the steps on its own, uses multiple programs, and fixes its own mistakes without assistance.

This technical exposure directly affects roles that were once well protected. Executives, legal professionals, software developers, and consultants are seeing parts of their daily work automated by software. At the same time, the job market is showing targeted signs of vulnerability. The unemployment rate for 15-to-24-year-olds in France rose from 19.1% in Q1 2025 to 21.6% in Q2 2026, according to INSEE. While this increase is linked to broader economic conditions, it raises questions about entry-level hiring in sectors where foundational tasks are now handed over to algorithms.

Understanding it

Being exposed to a technology does not mean losing one’s job.

Saying a job is 16.3% exposed merely means that a portion of its daily tasks can be performed by a machine. An accountant already uses spreadsheets without losing their job. The tool eliminates repetitive actions to free up time for advisory work, reshaping job content without eliminating employment contracts.

Déjà abordé sur Scénario IA : le pari à 725 milliards de la tech américaine (11 septembre)

To cushion the impact and prevent social divides, public authorities are attempting to organize a response through continuing education. On September 22, 2026, the French government launched the national “Impulsion IA 2027” plan. Backed by €80 million, this public initiative sets the ambitious goal of training 30 million workers in digital tools. The state aims to deploy a free learning platform called “Explore IA” and integrate certifications into apprentice training centers starting in 2027. However, the plan’s success remains uncertain given the scale of the task and the speed of technical developments.

The future of work will hinge on how quickly companies actually adopt these tools and how effectively workers can upskill. Three distinct scenarios emerge for the next two to five years. A smooth adaptation would transform occupations without net job losses, driven by new productivity gains. An uneven transition would concentrate losses in specific skilled sectors while creating positions elsewhere. Finally, a rapid displacement scenario would see automation destroy jobs faster than the economy can create them.

Share of French employment exposed to agentic AI (Senate estimate) 16.3% (5 million people) 2–5 year projection, Senate report April 2026
Unemployment rate among 15–24 year-olds in France 21.6% +2.5 points vs. Q1 2025 (19.1%)
Les chiffres

Le taux de chômage des 15-24 ans en France a progressé entre le premier trimestre 2025 et le deuxième trimestre 2026.

Source : INSEE, enquête Emploi (https://www.insee.fr/fr/statistiques/2532173)

Three Trajectories for Employment in the Face of Artificial Intelligence

What we're assessing
  • FavorableL'IA transforme les métiers sans détruire d'emplois nets grâce aux gains d'efficacité.
  • StableLes suppressions de postes restent localisées dans le conseil et l'ingénierie sans déséquilibre global.
  • DégradéL'automatisation rapide supprime des postes qualifiés plus vite que la création d'emplois.
Favorable
30%
Likely

AI transforms job content without large-scale job destruction

In this first scenario, companies integrate smart tools to assist their teams rather than replace workers. Employees delegate report drafting, data sorting, and document research to autonomous software. These time savings allow workers to focus on higher-value relational, creative, or strategic tasks. Companies see overall productivity gains, stimulating economic activity and driving hiring in consulting, cybersecurity, and technical project management.

Unlike sudden disruption forecasts, this path relies on a gradual rollout slowed by security regulations and infrastructure costs. The public “Impulsion IA 2027” plan enables rapid upskilling of the workforce, preventing a skills gap. The economy would then absorb this technological wave much like previous IT revolutions, without triggering a lasting surge in widespread unemployment.


Indicators affected
  • Part de l'emploi français exposé à l'IA agentique 16,3 % → Tâches absorbées sans perte de postes
  • Taux de chômage des 15-24 ans en France 19,5 % ↓ Stabilisation puis recul vers le niveau de 2025
The France angleWorker upskilling preserves overall employment balance and boosts French corporate productivity. ↑ Rather favorable for France.
Stable
45%
Likely

Targeted job cuts emerge in tech and consulting

In this second scenario, the impact of artificial intelligence fragments across sectors and skill levels. Basic software development, junior consulting, translation, and administrative roles face direct job cuts. Companies hire fewer recent graduates for simple execution tasks, as software agents perform this work faster and at minimal cost. In contrast, manual trades, care services, craft occupations, and industrial jobs experience no adverse impact.

Compared to the favorable scenario, overall macroeconomic balance remains stable but masks sharp individual and regional disparities. The June 2026 Trésor-Éco report already notes this lack of aggregate impact nationally, despite localized strains. Public training programs would ease difficulties for affected workers, but the job market could become durably harder to enter for young professionals with no experience.


Indicators affected
  • Part de l'emploi français exposé à l'IA agentique 16,3 % ↑ Tensions concentrées sur 5 millions d'actifs
  • Taux de chômage des 15-24 ans en France 22,0 % ↑ Maintien à un niveau élevé pour les jeunes diplômés
The France angleMandatory reskilling weighs on certain service sectors without destabilizing the broader national economy. ↓ Rather unfavorable for France.
Degraded
25%
Likely

Rapid automation drives net job losses in services

In this third scenario, autonomous agents improve much faster than companies and employees can adapt. Corporate management sharply cuts support, legal operations, accounting, and customer service teams to safeguard margins. Job eliminations far outpace the creation of new AI-related roles. Young professionals struggle to secure a first job as junior positions are automated or cut altogether as soon as they open.

Unlike the other two scenarios, the vocational training system and the state’s €80 million budget prove inadequate for the scale of the shock. Reskilling takes years, whereas software deployment takes only months. The unemployment rate for youth and intermediate professions would rise durably, triggering a major social crisis across knowledge-work sectors.


Indicators affected
  • Part de l'emploi français exposé à l'IA agentique 20,0 % ↑ Extension de l'exposition au-delà des 5 millions d'actifs
  • Taux de chômage des 15-24 ans en France 24,5 % ↑ Forte dégradation due au gel des embauches juniors
The France angleThe rapid elimination of skilled positions places a heavy burden on public finances and hinders entry-level hiring for graduates. ↓ Rather unfavorable for France.

Ordres de grandeur indicatifs pour les 3 scénarios ci-dessus, estimés avec l'information disponible à la publication et réévalués si la situation change — jamais des prévisions garanties. Learn more about our method → · How to read a probability →

Key takeaways

The emergence of autonomous artificial intelligence raises the question of net job destruction versus a simple transformation of working methods over the next two to five years.

The Senate report estimates that 16.3% of jobs in France—nearly 5 million workers—are directly exposed, while youth unemployment (15–24) reached 21.6% in Q2 2026.

Primarily a transformation: our most likely scenario (45%) anticipates localized losses in skilled services without an overall collapse of the labor market.

The practical rollout of the public Explore IA platform and training programs under the Impulsion IA 2027 plan in H1 2027 will test workers’ capacity to adapt.

Slightly negative

Our assessment of the impact for France: slightly negative.our baseline scenario of an uneven transition (45%) and the risk of net job losses (25%) outweigh a smooth adaptation (30%), pointing to real strains on skilled employment and youth hiring.

Si tu devais retenir 1 chose

En France, 16,3 % de l'emploi, soit près de 5 millions de personnes, est exposé à l'IA agentique d'ici deux à cinq ans selon le rapport du Sénat.

Quick glossary

Generative AI
Computer technology capable of creating text, code, images, or audio from prompts written in natural language.
Agentic AI
An artificial intelligence system capable of autonomously executing a complex series of tasks to achieve a designated goal without step-by-step guidance.
Trésor-Éco
Economic analysis publication produced by the Directorate General of the Treasury to inform public policy for the French Ministry of the Economy.
See all terms explained so far → Glossary

Sources

See also today's press roundup →

Vote before you see the outcome — follow us everywhere

Every day, a poll on our Telegram channel: vote for the scenario you think is most likely before you even see our probabilities.

Follow us also across all our channels: