AI cannot replace auditor judgement, flags automation bias in audits: NFRA

The audit regulator's first technology guidance says AI can support audit quality but cannot substitute professional judgement, scepticism or auditor accountability

Technology cannot be expected to exercise professional scepticism on the auditor's behalf and, while it can inform their professional judgement, auditors cannot invoke tech tools to explain away an inappropriate conclusion, the National Financial Reporting Authority (NFRA) said on Friday.

 

In its first paper on "Staff Series on Technology in Audit", NFRA laid down a principles-based framework for the adoption and governance of technology, including artificial intelligence (AI), generative AI and agentic AI, in statutory audits.

 

"The responsibility for the audit opinion remains with the auditor irrespective of the sophistication of the tools used," NFRA said in its paper titled "General Principles for Technology Adoption in Audit".

 

The guidance comes amid increasing use of automated tools and data analytics in audits and is intended to ensure that technology enhances audit quality without undermining professional scepticism or auditor accountability.

 

"Fluent and well-presented machine output is likely to invite automation bias, which is the tendency to accept plausible-looking results with less challenge than a human-prepared equivalent would receive. Engagement teams and reviewers must consciously counter this tendency," NFRA highlighted.

 

The financial reporting authority has stressed that adopting technology requires auditors and audit firms to make identifiable changes to how they govern quality, document work, train staff and monitor performance.

 

The regulator said that audit firms should establish formal processes for tool validation, monitoring and re-approval, strengthen data protection safeguards, and maintain transparent documentation of technology-assisted audit work. "Where technology use is material to the audit approach, the auditor needs to be able to explain such use," the authority added.

 

NFRA said that an audit firm should have quality control measures to treat technology as a resource whose risks are identified and addressed before deployment. The paper added that auditors must monitor the tool's performance throughout its working life and re-approve it when warranted, rather than relying indefinitely on an initial approval.

 

While acknowledging that technology is providing genuine gains in audit efficiency, it also introduces risks that traditional audit methodology was not designed to address. These include automation bias, opaque or non-deterministic outputs, model drift, data privacy exposure, and the risk that reliance on technology may narrow rather than deepen professional scepticism and efficiency.

 

Stressing the importance of data protection while using tech tools, NFRA said that a firm cannot rely on a vendor's terms of service or the convenience of a tool as a substitute for its own responsibility to protect client and personal information.

 

"NFRA has rightly ensured that accountability remains human. More importantly, it sends a clear signal that AI is not a convenient alibi for weak judgments or deficient audits. Auditors can leverage technology to work smarter, but they cannot outsource their judgment or accountability," said Jaspreet Bedi, Senior Partner – Audit & Assurance, Nangia & Co LLP.