Accountability remains the constant in a rapidly shifting landscape
The deployment of generative and agentic AI across audit engagements is continuing to accelerate at pace. The FRC published guidance – the Generative and Agentic AI Guidance – described as ‘the first from any audit regulator globally’, focused on identifying and mitigating risks regarding generative and agentic AI use in audit engagements, issued in March 2026 (the March Guidance).
Sitting alongside the FRC's guidance is the CCAB's draft statement to the profession on the ethical use of AI (the CCAB Statement), issued in June 2026 and currently open for consultation.
Both documents establish key regulatory expectations as AI adoption deepens in the accountancy sector. They also highlight that – ultimately – accountability still remains with professional accountants.
The March Guidance
The March Guidance organises AI-related risks to audit quality into three distinct categories:
- Deficient output risks exist where an AI tool produces a flawed output which is nonetheless relied upon in the audit. Root causes may lie in the inputs or the system's own performance, for example hallucinations (fabricated material), omissions, distortions, or faulty reasoning.
- Misuse of output risks occurs where the AI tool itself performs correctly, but the output is misinterpreted or misunderstood by the user, leading to inappropriate reliance.
- Risks associated with non-compliant methodology arise where a firm’s methodology in the use of AI tools permit approaches that fail to meet auditing standards.
The March Guidance is clear that such risks should be mitigated with reference to the specific tool and use case in question. When designing or testing tools, firms should consider the need for safeguards in system development, as well as whether there are appropriate approval and certification processes in place. Firms should also have robust training and governance arrangements, with all employees knowing how and when to use a specific AI tool.
Whatever the AI tool deployed, the March Guidance reaffirms that accountability remains paramount. In line with ISQM (UK) 1 and ISA (UK) 220, firms and the engagement partner remain fully responsible for audit quality. The need to have human review and oversight therefore cannot be overstated.
The March Guidance is best understood not merely as a codification of existing good practice, but, as the FRC itself signals, a ‘conceptual foundation for future FRC work in this area’. Firms are therefore well-advised to consider alignment with its various guardrails as a compliance priority.
The CCAB Statement
Unsurprisingly, the core message in the recently issued CCAB Statement is consistent with the March Guidance: AI tools do not diminish the personal responsibility of the professional accountant.
The CCAB Statement notes that professional accountants play a ‘key role’ in supporting responsible AI through governance, risk management, assurance, documentation and ethical leadership. The CCAB Statement also specifically applies the fundamental ethical principles to AI use and identifies a number of threats to such principles.
Examples include a self-interest threat from relying on an AI tool to reduce costs at the expense of quality, or a self-review threat from using an AI system that a firm itself was involved in designing.
The CCAB Statement helpfully concludes with some practical tips. This includes the need to be transparent with clients about how AI is used, validating all inputs and outputs, staying informed of best practice, and only using enterprise-grade tools which operate in a secure environment so that data privacy is not compromised.
ACCA members are strongly encouraged to engage with the CCAB consultation and to review the accompanying CCAB case studies in the ethical use of AI, which provide practical, scenario-based guidance directly relevant to accountancy practice.
Julie Matheson, Partner, Kingsley Napley
Ian Ko, Senior Associate, Kingsley Napley
Further reading
