Why AI raises the stakes for auditors
As artificial intelligence (AI) takes on a larger role in preparing and analysing company information, the need for human judgement, professional scepticism and quality-assured data grows with it. That was the message shared by XBRL International CEO John Turner in an interview with Tidningen Balans, the magazine of Swedish professional body FAR, before a meeting of regulators, companies and capital market representatives brought together for XBRL Sweden’s annual conference in Stockholm, to discuss the future of digital reporting.
AI puts a premium on skills that auditors already have: scepticism, and the ability to follow information through processes and controls. John expects audit teams to bring in more specialists in statistics, and expand existing skills in data analysis and modelling, while human judgement and dialogue with management continue to be vital. Responsibility stays with people too. “You can’t stand before a regulator and say that you trusted the machine.”
The interview tackled a key question about AI: if large language models can read annual reports directly, does structured data still matter? John explains that AI gets markedly better results with structured than with unstructured inputs, and with financial information, where quality is critical, users need to be able to trust what they get.
When a company publishes one clear digital version of its report, every user starts from the same facts. Point thousands of different AI systems at the same PDF and you get thousands of private versions of the truth. As more reporting is consumed by machines, the digital representation of a company’s accounts becomes as important as the traditional report, and over time, the auditor’s role in vouching for it will grow accordingly.
Read the full interview here (in Swedish).
