ICAEW and ACCA Guidance on AI for Accountants

Q: What do the two bodies agree on?

ExpectationICAEWACCA Member stays responsible for the workCode applies in full, scepticism and judgement apply to all AI useMembers cannot outsource scepticism and judgement to technology ConfidentialityUploading client data to AI tools flagged as a key riskClient data in public tools likely to breach confidentiality without consent Automation biasNamed in the Code of Ethics 2026Undue reliance on technology named under objectivity Firm policyNot set out as a separate requirement in the CodeAll firms should have firmwide AI policies Telling clientsNot set out in the CodeClients should ideally be told which tools are used TrainingFree ethics course updated with AI units in January 2025Firms should review training and CPD to include AI

Q: What does this look like on a real job?

Illustration only. A two partner ACCA firm uses a general purpose chatbot to draft a letter explaining a client's capital gains position. A junior pastes in the client's name, the sale price and the base cost. The draft cites a relief that does not apply to the asset sold. Measured against the guidance: putting identifiable client data into a public tool without consent is likely to breach confidentiality; the output needed checking against current law before it went anywhere; and the partner who signs the letter is responsible for it whether or not AI was involved. A firm following the guidance would have anonymised the facts, used an approved tool, had a qualified reviewer check the relief and kept a note of the tool and prompt on file.

ICAEW and ACCA have not written a separate AI rulebook. Both apply their existing Codes of Ethics to AI, and both now publish practical material on top: ICAEW through its Code, its Ethics and AI guide and its generative AI resources, ACCA through its Code, a technical article on AI risk and topical guidance on applying PCRT to AI tools in tax work. The common thread is that the member stays responsible for any work an AI tool helps produce.

Tax investigation insurance from £108.75 a year (£9.06 a month).Pays up to £100,000 of your accountant's fees if HMRC opens an enquiry. No excess. Buy online in a few minutes.

See my price in 2 questions How the cover works Accountant? Offer cover to your clients or call 01795 841 310

Do ICAEW and ACCA have specific rules on generative AI?

Neither body has a standalone AI regulation that members must sign up to. What they do is apply the fundamental principles that already bind members, and publish guidance on how those principles work when a tool such as a chatbot or an AI drafting assistant is part of the job. ACCA says its topical guidance on AI in tax work derives its requirements from the fundamental principles in PCRT and the Standards for Tax Planning, and that a failure to follow them may lead to the disciplinary process.

This page sets out what each body has published. For a wider discussion of the ethical questions AI raises in day to day work, see the ethics of AI use in practice.

Which parts of the ICAEW Code of Ethics apply to AI?

The ICAEW Code of Ethics lists five fundamental principles: integrity, objectivity, professional competence and due care, confidentiality and professional behaviour. ICAEW updated its Code to align with the international code issued by IESBA, and the updated Code was enforced from 1 July 2025. That update covered three areas, one of which was technology.

The current edition, the ICAEW Code of Ethics 2026, applies from 1 July 2026. It names automation bias, described as a tendency to favour output generated from automated systems, as something that can get in the way of a member acting properly. It also says, in the context of threats from technology, that the data available might not be sufficient for the effective use of the technology, and that the level of oversight and internal controls over a tool affects how serious a threat is. Professional competence and due care includes keeping up with technology related developments relevant to the member’s work.

Worried about an HMRC check? Tax investigation insurance pays your accountant’s fees if HMRC opens an enquiry or compliance check. Solar Protect cover pays up to £100,000 with no excess. It never pays tax, interest or penalties.

From £108.75 a year (£9.06 a month) for individuals, £117.57 for sole traders and £207.50 for companies and partnerships, including IPT and the £20 admin fee.

See my price in 2 questions or call 01795 841 310.

★★★★★ Rated 4.9 out of 5 on Trustpilot from over 160 reviews

What else has ICAEW published on AI?

ICAEW’s Trust and Ethics team published an Ethics and Artificial Intelligence guide on 18 October 2024, drawn from roundtables with more than 40 participants from practice, industry, policy and academia. Its main points for practitioners are:

professional scepticism and professional judgement should apply to every AI interaction
automation bias, the human tendency to trust and prefer recommendations from automated systems, is a risk to counteract
confidentiality is a key pillar of ethical practice, and uploading client data to AI tools is a specific risk
staff need the skills to use AI and to evaluate its output critically

ICAEW also runs a generative AI guide with a section on the ethical points to keep in mind, and in January 2025 it added two units to its free ethics elearning course for members, covering the ethical considerations of using AI and building an ethical culture.

What does ACCA say about AI risk in a practice?

ACCA’s UK technical team published an article on the key areas of risk when adopting AI in March 2026. It is written for firms and for ACCA members and regulated individuals within them, and it is direct:

all firms should have firmwide policies covering responsible use of AI tools, including limits such as bias and hallucinations and how to escalate adverse incidents
one of the major risks of unchecked AI use is compromising confidential client data
clients should ideally be made aware of any AI tools that will be used, and the engagement letter can explain how far the firm relies on AI output
members cannot abdicate or outsource their responsibilities for professional scepticism and judgement to technology
firms should review their training and CPD programmes to build in AI training
before buying a tool, check what data it collects, where it is stored, how long it is kept and whether it trains on customer data

The ACCA Code of Ethics and Conduct, restructured on 1 July 2025, lists the same five fundamental principles. Its description of objectivity refers to undue influence of, or undue reliance on, individuals, organisations, technology or other factors.

ACCA has also published, with the Irish accountancy bodies in CCAB Ireland, Technical Alert 01/2025, pointers for a firm’s AI policy, dated February 2025. It is aimed at the Irish market but its checklist is useful anywhere: a human in the loop for critical decisions, only approved tools, no data that would breach confidentiality, and independent verification and fact checking of outcomes. The page on an AI policy for an accountancy practice turns that into a working document.

How does PCRT apply when AI is used for tax work?

Tax work carries its own layer. ACCA’s topical guidance on the application of PCRT to the ethical use of AI tools is dated 19 January 2026, and ICAEW publishes topical guidance on the same subject. The ACCA document says members are ultimately responsible for any work they produce, regardless of whether AI has been involved, and suggests treating AI output as the work of a less experienced junior colleague.

In practical terms it expects members to check for hallucinations such as case law that does not exist, confirm that outputs reflect the client’s circumstances and current law, and be able to explain how a conclusion was reached. It says client data put into publicly available AI tools is likely to breach confidentiality unless the client consents, and that inputs to public tools should be anonymised and generic. It suggests a statement in the engagement letter that AI enabled software may be used, and keeping an audit trail of the tools used.

What do the two bodies agree on?

Expectation	ICAEW	ACCA
Member stays responsible for the work	Code applies in full, scepticism and judgement apply to all AI use	Members cannot outsource scepticism and judgement to technology
Confidentiality	Uploading client data to AI tools flagged as a key risk	Client data in public tools likely to breach confidentiality without consent
Automation bias	Named in the Code of Ethics 2026	Undue reliance on technology named under objectivity
Firm policy	Not set out as a separate requirement in the Code	All firms should have firmwide AI policies
Telling clients	Not set out in the Code	Clients should ideally be told which tools are used
Training	Free ethics course updated with AI units in January 2025	Firms should review training and CPD to include AI

What should a practice do with this guidance?

Read together, the guidance points to a short list of things a reviewer from either body would expect to see:

A written AI policy that names the approved tools and what data may go into each.
A review step for every AI output that leaves the firm, done by someone competent to spot errors.
Wording in the engagement letter that tells clients AI tools may be used. See AI clauses in engagement letters.
A record of which tool was used on which job, with prompts kept where the output shaped advice.
Training logged as part of each person’s CPD.

None of this changes who carries the risk when an error gets through. A claim for a mistake in AI assisted work is a claim against the firm, which is why some practices check how their PI policy responds and whether AI insurance for accountants is needed alongside it.

What does this look like on a real job?

Illustration only. A two partner ACCA firm uses a general purpose chatbot to draft a letter explaining a client’s capital gains position. A junior pastes in the client’s name, the sale price and the base cost. The draft cites a relief that does not apply to the asset sold.

Measured against the guidance: putting identifiable client data into a public tool without consent is likely to breach confidentiality; the output needed checking against current law before it went anywhere; and the partner who signs the letter is responsible for it whether or not AI was involved. A firm following the guidance would have anonymised the facts, used an approved tool, had a qualified reviewer check the relief and kept a note of the tool and prompt on file.

Frequently asked questions

Is the ICAEW or ACCA AI guidance mandatory?

The Codes of Ethics are binding on members, and both apply to work done with AI. The supporting guides explain how to meet the Codes. ACCA says failure to follow its PCRT AI guidance may lead to the disciplinary process.

Can I put client data into ChatGPT or a similar tool?

ACCA’s PCRT AI guidance says client data put into publicly available tools is likely to breach confidentiality unless the client consents, and that inputs should be anonymised and generic. The page on AI tools and client data covers the data protection side.

Do I have to tell clients I use AI?

ACCA says clients should ideally be made aware of the AI tools to be used, and its PCRT guidance suggests a statement in the engagement letter that AI enabled software may be used.

Does AI training count towards CPD?

ACCA expects firms to review training and CPD programmes to include AI, and ICAEW’s free ethics course now includes AI units.

Using AI in your practice? Solar AI Cover brings professional indemnity, cyber and liability cover together for firms that use or build AI, and Solar Protect runs fee protection schemes for accountancy practices. See AI insurance for accountants or ask about a practice scheme.

Sources

Reviewed by Christine Anderson, Broker, Solar Protect. Last reviewed 6 November 2026. Part of Accountapedia, the Solar Protect guide to UK tax and accounting. This page is general information, not personalised advice.