Jeff Raikes says AI talent debt is appearing in workforce data

Jeff Raikes says his AI talent debt warning is now showing up in research on work and education. The former Microsoft executive and former Gates Foundation CEO argues that companies using AI to take over junior-level tasks risk weakening the human judgment they will need later.

In a Fortune commentary, Raikes pointed to Microsoft’s 2026 Work Trend Index, which surveyed 20,000 workers in 10 countries. Microsoft found that the most valuable human abilities in an AI-heavy workplace are critical thinking and the capacity to assess AI output, rather than technical skills alone.

According to Microsoft, only 16% of workers have developed the judgment to shift smoothly between directing AI systems and doing tasks themselves. Microsoft calls that group “Frontier Professionals,” and Raikes highlighted the company’s finding that they sometimes complete work without AI to keep their own thinking sharp.

What is AI talent debt?

Raikes uses “AI talent debt” to describe the long-term cost companies may build up when they let AI replace the entry-level work where employees learn judgment, problem-solving and professional habits. The concern is that short-term efficiency could leave employers with fewer people able to supervise AI, catch errors and take responsibility for decisions.

Raikes said other research points in the same direction. Gartner has predicted that through 2026, concern about declining critical-thinking skills will lead half of global organizations to require assessments that do not allow AI. A RAND Corporation study published in the spring found that most students using AI for homework worry the tools are hurting their ability to think independently.

The debate reaches beyond corporate training, Raikes argued. He wrote that the skills employers now value in an AI workplace overlap with the skills people need to function as citizens, including reasoning, discernment and the ability to challenge information rather than accept it.

How does Raikes say colleges should respond?

Raikes said many colleges are treating AI mainly as a way to make credentials cheaper, faster and more job-focused. He cited a Pew Research Center survey that found 70% of Americans believe U.S. higher education is headed in the wrong direction, creating pressure on institutions to show practical value.

Raikes also cited higher education leader Paul LeBlanc, the former head of Southern New Hampshire University, who argues in a forthcoming book, “Reclaiming Purpose: The University in an AI World,” that education should use AI to strengthen the human capacities that are hardest to automate. Raikes described those capacities as relationship-building, collaboration, ethical reasoning and judgment.

Raikes said the institutions most important to the future workforce are often outside the center of national AI debates. He cited Community College Research Center data that community colleges enroll roughly 40% of U.S. undergraduates, and he grouped them with HBCUs and regional public universities as key training grounds for first-generation students, working adults and other students for whom college can be financially difficult.

Some of those schools are already adding AI programs, according to examples Raikes cited. Community colleges are working with local employers on applied AI programs, HBCUs are building training partnerships with Google, NVIDIA and IBM, and Illinois has advanced legislation that would allow community colleges to offer bachelor’s degrees in high-demand fields.

Raikes said those efforts often focus on AI competency, such as prompting, summarizing and using generative tools for analysis. He argued that business leaders should also fund mentored learning that builds durable judgment, especially at community colleges and HBCUs, because AI literacy without critical thinking will not be enough for the workforce companies say they need.

This story draws on original reporting from Fortune.