AI needs better data, not bigger promises
As regulators around the world wrestle with how artificial intelligence can strengthen oversight of corporate reporting, XBRL US has delivered a clear statement to the Public Company Accounting Oversight Board (PCAOB): the future of audit supervision may depend less on flashy AI models and more on the quality of the data fed into them.
In a response to the PCAOB’s consultation on its forthcoming 2026–2030 strategic plan, XBRL US urged the audit watchdog to make greater use of the vast pool of structured financial data already available through US Securities and Exchange Commission filings. The submission argues that AI systems perform best when working with standardised, machine-readable information, which is precisely the kind of data produced through XBRL-tagged corporate reporting.
For the PCAOB, whose mandate centres on protecting investors and improving audit quality, XBRL US suggests the opportunity is immediate and concrete: use AI tools against structured SEC filing data to identify anomalies, detect potential non-compliance with US GAAP requirements, and analyse disclosures at scale.
The letter envisions an AI-driven supervisory model, beyond the traditional audit toolbox, whereby automated rules and AI-enabled analysis could help the PCAOB review thousands of filings rapidly and consistently, highlighting unusual reporting patterns or weaknesses that might otherwise escape attention.
For years, advocates of structured data have argued that digital reporting formats such as XBRL would eventually unlock more powerful analytics and regulatory oversight. The rapid emergence of generative AI and large-scale machine learning has sharpened that argument. AI systems are only as reliable as the data they consume, and unstructured disclosures remain expensive and error-prone to analyse.
The PCAOB’s strategic plan will guide the organisation’s priorities and budgeting through 2030, making the consultation a significant opportunity for stakeholders seeking to shape how the regulator approaches technology and data in the years ahead.
Read the letter here.
