Week 8: The Professions Most Exposed to AI Disruption in 2025

Artificial intelligence is no longer just a helpful tool—it’s changing the very shape of the job market. A new study from OpenAI, called GDPval, examined the performance of AI models on real-world tasks across 44 professions in the nine largest sectors of the U.S. economy. The results indicate which jobs—and corresponding salary levels—could be most impacted in the years ahead.

The Jobs Most at Risk

The first wave of impact is hitting knowledge workers—people whose jobs revolve around digital documents, analysis, and communication. These are often high-wage professions, meaning the economic stakes are large.

• Finance and Professional Services: Roles like financial analysts, accountants, consultants, and lawyers, typically earning between $80,000 and $160,000+, were found to be highly exposed. AI can now produce spreadsheets, financial models, legal memos, and strategy reports at a quality that rivals seasoned professionals (OpenAI, 2025).

• Information and Media: Journalists, editors, and marketing professionals, with wages in the $60,000–$120,000 range, are also vulnerable to exploitation. AI shows strong performance in creating and formatting content, from news briefs to marketing decks (OpenAI, 2025).

• Healthcare Administration: While doctors and nurses aren’t being replaced, administrative tasks like case documentation, summaries, and reporting—jobs that pay around $70,000–$110,000—are increasingly within AI’s reach. The caveat is that human oversight remains crucial to ensure accuracy and safety (OpenAI, 2025).

• Retail, Wholesale, and Strategy: Executives, sales analysts, and account managers, with salaries from $70,000–$140,000, rely on forecasts and presentations—deliverables that AI can generate quickly, if sometimes superficially (OpenAI, 2025).

• Engineering and Manufacturing (Digital Work): Industrial engineers and CAD designers, who often earn $80,000–$120,000, already utilize digital tools extensively. AI can speed up design iterations but usually struggles with context and formatting (OpenAI, 2025).

• Government and Real Estate: Policy analysts, program managers, and brokers, with salaries in the $60,000–$100,000 range, are also seeing AI encroach on their report writing and analysis tasks (OpenAI, 2025).

Together, these professions represent over $3 trillion in wages annually in the United States.

Scale and Scope of Impact

The study highlights just how significant this change could be. On average, the professional tasks tested required seven hours of human effort and were valued at $300–$400 each. Yet, AI systems—when paired with expert oversight—completed them faster and more cost-effectively, and in many cases, matched human quality (OpenAI, 2025).

The implications are wide-ranging:

• Productivity gains: Experts who “try AI first” before doing tasks themselves save measurable time and money.

• Displacement risks: High-paying, digital-heavy jobs are under the most pressure, not just routine office work.

• Societal scope: Because affected roles span finance, healthcare, government, and media, the ripple effects touch decision-making, trust in institutions, and economic stability.

A Shift in What Work Means

What makes GDPval different from past studies is its realism. Instead of testing abstract reasoning, it evaluated actual deliverables—PowerPoint decks, spreadsheets, CAD drawings, financial models, and policy briefs. These tasks were designed by industry experts with an average of 14 years of experience. The fact that AI can now match or exceed some of their work underscores how quickly the landscape is shifting (OpenAI, 2025).

While AI is not ready to replace professionals outright—it still makes formatting mistakes, misses context, and occasionally hallucinates information—the trajectory is clear. Each new model closes the gap further, suggesting that oversight and integration, rather than replacement, will be the near-term reality.

The Bottom Line

The most exposed professions aren’t in factories or low-skill offices—they’re in finance, law, healthcare, media, government, and engineering. These jobs form the backbone of the service economy and carry some of the highest wages. As AI models improve, the question is no longer whether these professions will be impacted, but how workers, employers, and society will adapt.

---

References

OpenAI. (2025). GDPval: Evaluating AI model performance on real-world economically valuable tasks. Retrieved from https://evals.openai.com

Brynjolfsson, E., Li, D., & Raymond, L. (2025). Generative AI at work. The Quarterly Journal of Economics, 140(2), 889–942.

Chen, W. X., Srinivasan, S., & Zakerinia, S. (2025). Displacement or complementarity?: The labor market impact of generative AI. Working Paper.