Which jobs will become automated, reorganized, grow, or evolve less than expected?

We talk about jobs "exposed" to AI. That's a starting point, but it's insufficient. The real question is: Which jobs will become automated, reorganized, grow, or evolve less than expected?

1️⃣ Exposure to AI is not enough. Knowing that AI can perform certain tasks doesn't mean a job will actually be transformed. There's a world of difference between technical capability and employment impact: adoption, organization, trust, regulation, uses…

2️⃣ We need to consider four variables, not just one. The model proposed here combines: AI exposure, human need, demand elasticity, and actual usage. In other words, we move from technical possibility to economic and social relevance.

3️⃣ Not all the professions exposed will evolve in the same way. The report distinguishes 4 main archetypes:
– Jobs at high risk of automation (18%)
– Those who will primarily reorganize (24%)
– Those that could grow thanks to AI (12%)
– Those with little immediate change (46%)

4️⃣ Humans remain central to many professions. Even as AI advances, some activities remain difficult to replace for three reasons: responsibility, relationships, and physical presence. This explains why professions like teacher, nurse, or lawyer can be highly exposed… without being replaceable.

5️⃣ More productivity doesn't mean fewer jobs. If AI lowers the cost of a service, demand can increase: more access, more volume, more frequency, more personalization. In some cases, AI doesn't compress activity: it expands it.

6️⃣ There's a huge gap between what AI can do… and what's actually being used. The report highlights a significant gap between AI's theoretical capabilities and its real-world use in organizations. The labor market isn't keeping pace with technological progress. Paradoxically, the category of "Highest Risk of Automation" shows the greatest underutilization of AI at present.

7️⃣ For decision-makers, the right answer varies depending on the situation. Depending on the business, it may be necessary to: monitor and support transitions, redesign tasks and roles, invest to capture growth opportunities, or simply measure more effectively and continue observing. The real challenge is not just “adopting AI,” but adopting the right strategy for each business function.

8️⃣ The final strategy In the coming months, the organizations that make the best decisions will be those with the most accurate reading of the actual work of the teams, and will thus know what to automate, what to reorganize, what to develop and what to leave human.

And you? Do you feel this gap between what AI could do… and what you actually entrust to it on a daily basis? 👇

Source: OpenAI Economic Research study: https://cdn.openai.com/pdf/the-ai-jobs-transition-framework_report.pdf

To delve deeper: this occupational analysis echoes our article on the AI agents in HR and to our survey on the impact of AI on employment.