AI in finance and accounting
Objectivity is defined as exercising judgement without being compromised by bias, conflict of interest, or undue influence of, or undue reliance on, individuals, organisations, technology or other factors. The word technology sits inside the definition of a fundamental principle, which makes over-reliance on a model a breach of objectivity rather than of care.
§ 1 — Who it binds
Every professional accountant, in business and in public practice, in every professional or business judgement. Objectivity is a fundamental principle rather than an engagement-level requirement, so it is engaged wherever technology output informs a judgement and not only where independence is in issue. The Code takes effect through each IFAC member body that has adopted it; US CPAs are governed by a separate instrument this page does not characterise.
§ 2 — In practice
Section 110 sets out the five fundamental principles, and objectivity is defined at 110.1 A1(b) as exercising professional or business judgement without being compromised by (i) bias, (ii) conflict of interest, or (iii) “undue influence of, or undue reliance on, individuals, organisations, technology or other factors”. The third limb is the one this page turns on. Most writing that reaches the subject at all treats over-reliance on a model as a competence problem — the accountant did not check enough — and the Code does not classify it that way. It classifies it as a compromise of objectivity, which is a different kind of failure: not a shortfall of skill or effort, but a judgement that was not the accountant’s own.
The distinction has consequences an assessor can use. A competence failure is answered with evidence of work: more testing, a larger sample, a second reviewer. A compromise of objectivity is answered with evidence of independence of mind, which is a different artefact — what the accountant would have concluded had the tool not produced an answer first, and whether anything in the process gave them the chance to form that view. That is why the order of operations matters so much here. Forming an expectation before seeing the output is a control against limb (iii); checking the output afterwards is not, however carefully it is done. It is also why the Code names automation bias separately: bias is limb (i) of the same definition, and over-reliance is limb (iii).
The wording is older than the technology revisions — it came in with Role and Mindset — so the Code has held over-reliance on technology to compromise a fundamental principle since before the current tools were released. What the definition does not do is set a threshold. It says undue, which is a judgement about circumstances rather than a quantity of checking, and undueness rises as the decision depends more heavily on the output and as the accountant is able to find out less about how it was produced. The factors bearing on that assessment are not in Section 110 at all. They are in R220.8, which is why the two provisions have to be read together and are almost never cited together.
§ 3 — What a weak answer looks like
Objectivity treated as a conflicts question. The function’s answer is that nobody holds a financial interest in the tool, the supplier was selected competitively, and no one’s bonus turns on the output — all of which addresses limb (ii) and none of which addresses limb (iii). Conflict of interest and undue reliance sit as separate limbs of one definition precisely because they fail differently: one is about what the accountant wants the answer to be, the other about whether they arrived at it themselves.
§ 4 — What discharges it
The artefacts an assessor asks to see, and what makes each one sufficient rather than merely present.
01
What the accountant expected, on what basis, written first. It is the only artefact capable of distinguishing a conclusion they reached from one they were handed and did not disturb.
02
Self-reported accuracy is a claim by the party whose work is under review. What counts is whether anything independent corroborated it, and where nothing did, whether that limited the conclusion drawn.
03
Occasions on which a person departed from the output, with reasons. A period in which nobody ever departed is not proof of a good tool; it is the pattern the third limb describes.
04
Why this degree of dependence was appropriate here, given what could be established about how the output was produced. Undueness is circumstantial, so the record has to be circumstantial too.
§ 5 — Worked example
A group reporting manager reviews an AI tool’s classification of 40,000 supplier invoices into expense categories for segment reporting. The tool reports 97 per cent confidence. The manager samples fifty, finds two misclassifications, corrects them and accepts the rest. Asked later how the population was concluded acceptable, the answer is that the sample rate matched prior years and the errors found were immaterial.
Is this a competence question or an objectivity question?
Both, and separating them is the useful part. Whether fifty was a defensible sample is the competence question, the manager has answered it, and the answer is not obviously wrong. The objectivity question is untouched. Nothing in the account describes a view formed independently of the tool: no expectation of what the segment split should look like, no comparison against the prior year’s mix, no attention to whether the two errors were random or clustered in the categories this kind of tool finds hardest. The 97 per cent is the tool’s own account of itself, and it has been treated as an input to the judgement rather than as a claim wanting corroboration. Under the Code that is limb (iii) territory, and the remedy is not a bigger sample. It is a step that produces an expectation the output can then be tested against — usually cheaper than doubling the sample, and the only thing that shows the conclusion was the accountant’s.
§ 6 — Elsewhere
Where another instrument addresses the same obligation. These are correspondences, not comparisons — the Council does not rank one framework against another.
Finance and accounting standards
120.12 A2 names the tendency to prefer automated output. The objectivity definition places over-reliance on technology inside the principle itself, so one episode can engage both limbs of the same definition.
Finance and accounting standards
Section 110 states the principle that over-reliance compromises. R220.8 supplies the factors bearing on whether a particular reliance was appropriate, which is where the assessment actually happens.
A correspondence indicates that two instruments address the same underlying obligation. It is not a mapping endorsed by either body, not a statement that one satisfies the other, and not a judgement about which is more demanding.
§ 7 — When it applies
29 August 2024
Publication date of the 2024 edition of the IESBA Handbook, shown as an edition marker. The reference to technology in the objectivity definition came with the Role and Mindset revisions, whose IESBA commencement date has not been verified. Not a compliance deadline.
§ 8 — Exposure
No Code fine; a compromised fundamental principle is a matter for the adopting member body
Objectivity is one of the five fundamental principles. A member body that has adopted the Code investigates alleged breaches and may reprimand, fine, impose conditions on practice, or withdraw membership and the practising certificate. IESBA imposes nothing and no regulator enforces the principle as such, so a person who is not a member of an adopting body is not reached at all.
§ 9 — Where this is assessed
Examined in one credential, in the domains named on each card.
Responsible use and accountability · 15% of the paper
For finance and accounting professionals who use AI in analysis, close and reporting work — and verify the output before the numbers are relied on.
§ 10 — provenance
This page sets out what the instrument requires and what discharges it. The official text is the authority — these go straight to it.
§ 11 — Also read
Certification
What this page sets out is examined, on work the candidate has not seen, in the credential beside it. Every paper is machine marked against its key and every result is reviewed and released by a named person, and every credential is on the public register.
AIAC-FNAF
Certified AI in Finance & Accounting — Fundamentals2 hours · remote, proctored · valid 2 years
Ten questions on using AI at work. About two minutes, free. You finish as one of six types, with a card to share.
A free quiz from the AI Assurance Council. Not a certification.