PwC is launching a critical-thinking and judgement certification programme for more than 85,000 partners and employees as artificial intelligence takes over tasks traditionally performed by junior professionals.
The initiative, announced on 9 November, aims to strengthen employees’ ability to evaluate information, question assumptions and exercise independent judgement when working with AI.
The programme will include personalised simulations requiring employees to challenge AI-generated conclusions. In one example, an auditor could be asked to review an AI-generated assessment of a company’s internal controls, identify deficiencies and question the tool’s findings.
The training will combine diagnostic assessments, masterclasses, practical exercises and workplace tools. PwC plans to begin rolling it out in December 2026 across its US and Mexico offices and offshore service centres, with completion targeted for June 2027.
The firm expects to invest more than $100 million in the initiative, including programme development and support for universities. It also plans to make its critical-thinking framework available to universities in early 2027.
The challenge of maintaining professional judgement
PwC’s initiative reflects a broader challenge confronting the Big Four: AI is increasingly performing the tasks through which junior auditors traditionally developed their professional judgement.
For generations, junior staff learned to recognise inconsistencies, question explanations and assess evidence by performing routine audit procedures under supervision. Repeated exposure to real engagements helped develop the judgement and professional scepticism needed for more complex work.
William Englehaupt, a former Arthur Andersen professional, recently examined how Andersen’s intensive training model offers lessons for the AI era. He argued that firms must find ways to replace the practical learning opportunities disappearing as AI takes over routine assignments. Formal training can teach technical foundations, but professional judgement has traditionally developed through experience.
KPMG US is confronting the same problem. Earlier this year, the firm redesigned its audit internship programme, introducing fraud-detection simulations and exercises requiring interns to investigate anomalies and make decisions using incomplete information.
Research conducted by KPMG and the University of Texas at Austin illustrates the difficulty. In a study involving 523 early-career professionals working with AI agents, 50.1% improved on the AI’s performance, while 25.8% produced broadly equivalent results and 24.1% performed worse than the AI working alone.
The findings suggest that access to sophisticated AI tools does not automatically improve professional judgement. Some employees accepted AI-generated conclusions with limited scrutiny, while others attempted to challenge the technology without identifying the most relevant weaknesses.
The risk of automation deference
The danger extends beyond inadequate training.
Wayne Banks, a chartered accountant and former global CFO, has warned about the risk of automation deference: the gradual erosion of professional scepticism when employees become accustomed to accepting polished, apparently credible AI-generated work.
An inexperienced auditor reviewing an AI-generated assessment might fail to question an unsupported assumption or overlook contradictory evidence because the conclusion appears technically sound.
Banks has argued that firms need documented verification procedures, clear accountability and controls requiring employees to test AI-generated conclusions rather than simply approve them.
PwC’s programme represents another attempt to address this challenge. The question is: can simulations and formal certification develop the same depth of professional judgement that previous generations acquired through years of supervised audit work?
As AI assumes more responsibility for preparing analyses and evaluating evidence, the ability to challenge its conclusions becomes increasingly important. Whether the Big Four can preserve that capability while fundamentally changing how auditors learn will have significant implications for audit quality.
Want to stay up to date on all things Big Four around the world?
Check out the News section and subscribe to Big4News for weekly deep dives and briefings.
About the author: Claudine Cassar is the founder and editor of Big4News, covering audit, consulting, regulation and governance across Deloitte, PwC, EY and KPMG.