By Olivia Hart

PwC is putting $100 million into a critical-thinking and judgment certification for more than 85,000 employees and partners after training workers to use new technology. The program turns AI adoption into scrutiny: staff will fact-check automated outputs, challenge assumptions and identify when human judgment should take over.

Bloomberg Tax reported the $100 million initiative as PwC’s response to a workplace increasingly dominated by AI. It formalizes a shift from teaching employees to operate new systems toward teaching them to evaluate what those systems produce.

What does the certification teach?

The certification treats evaluation as a repeatable process rather than a general awareness exercise. Its central subjects are automated outputs, the assumptions behind them and the point at which a person should resume control.

“Our goal is to give our people a practical, repeatable approach to evaluating information, challenging assumptions, and exercising sound judgment—particularly when AI is involved,” wrote Yolanda Seals-Coffield, PwC’s chief people and inclusion officer, in a press release. Her formulation puts judgment before AI, because the technology’s output has to withstand scrutiny rather than command acceptance.

How does the training work in practice?

Margaret Burke, PwC U.S.’s leader of talent acquisition and development, used an AI model’s tax analysis in an interview with Human Resources Director. Under the training, employees will:

• Trace the model’s logical process.
• Identify where the model pulls its data.
• Decide whether they feel confident in the conclusion or whether the model needs more fine-tuning.

The sequence leaves the employee responsible for the decision. A polished answer does not remove the need to inspect whether the reasoning and source material support it.

“It’s an opportunity to practice the judgment involved in making a decision, rather than simply learning about critical thinking in theory,” Burke said. That distinction makes the program a workplace judgment exercise, using tax analysis as its concrete example, rather than a theory module.

Why is PwC acting now?

PwC is also trying to counter a decline in critical thinking that concerns experts across several fields. A recent study found that higher AI usage was related to lower critical thinking through “cognitive offloading,” defined as transferring a skill to a machine and becoming weaker at it.

Familiar precedents include GPS use weakening some people’s sense of direction and smartphone use weakening their ability to remember phone numbers. The concern is that generative AI can offload more executive functioning than earlier technologies, making collective cognitive offloading a larger risk to how people think.

What does Farahany warn?

Nita Farahany, a philosophy professor, addressed that risk in an interview with the American Psychological Association. “One of the things that’s seductive but also troubling about generative AI is the ability to offload much more executive functioning than we have in the past,” she said.

Her second question defines the stakes for employers and professional teams: “What does it mean for society if humans are passively receiving information but no longer able to critically interrogate it?” AI’s expanding use could leave employees receiving information without independently questioning it, even when the systems generating it appear efficient.

What changes for practitioners?

For PwC’s workforce, the new standard is specific: inspect the logic, check the data and decide whether the answer merits confidence. The company is not asking employees to treat AI as useless; it is asking them to withhold trust until the output survives review.

That changes the practical role of technology users from answer recipients to active evaluators. For PwC, that standard applies directly to the workforce covered by the certification, with automated tax analysis providing the training example.

Burke’s emphasis on practicing judgment matters because the certification is framed around decisions rather than abstract theory. PwC’s $100 million program makes trust conditional on a completed review: accept the output, request fine-tuning or apply human judgment when the evidence is insufficient.