FCA authorised · FRN 724952 0117 325 0027Quote & buy →
Aipedia: Actuaries
Actuaries can use generative AI to draft reports and advice letters, but the signing actuary must ensure each one is accurate, complete and clear. TAS 100 and Code Principle 6 apply to every communication. A smooth summary that drops a caveat can mislead trustees, boards and third parties who rely on it.
Part of: AI in Actuaries
In short
Generative AI is good at turning technical results into readable prose, which is exactly where actuarial communication can go wrong. TAS 100 sets communication principles (P6 and P7), and the Actuaries' Code requires members to communicate appropriately. The IFoA warns about the veracity of third-party generative AI tools. Typical errors are missing limitations, overstated certainty, invented explanations and numbers that no longer match the model. Claims can come from the client and sometimes from third parties who relied on the report. Actuarial PI usually responds to misstatement claims subject to the policy terms, and engagement terms and reliance restrictions matter.
Actuarial reports carry results to people who make decisions: trustees, boards, investment committees, regulators and courts. Generative AI tools may be used to produce first drafts of valuation reports, advice letters, board papers and executive summaries, to rewrite technical text in plainer language, and to summarise long model documentation. The Lloyd's Market Association's discussion paper includes report generation among actuarial AI uses.
The benefit is speed and readability. The risk is that a well-written draft gives a false sense of completeness. A generative model writes what sounds right; it does not know which caveat the actuary considered essential, which sensitivity matters most to this client, or that the numbers were updated after the draft was produced.
No standard forbids AI drafting. The point is that responsibility for what the communication says stays with the actuary who issues it, and any client information put into a drafting tool must be handled confidentially; see client data confidentiality.
Illustrative scenario: an actuary uses a generative AI tool to turn a long funding valuation report into a summary for trustees. The summary is clear and accurate on the headline result but drops the paragraph explaining how sensitive the result is to the longevity assumption. The trustees share the summary with the sponsor and its lender, who treat the funding position as firmer than it was. When the next valuation shows a deterioration, the trustees ask why the sensitivity was not in front of them.
Other failure points: the AI invents a plausible reason for a change in results, uses figures from an earlier draft, describes an assumption inaccurately, or produces a confident tone that conflicts with the actuary's actual view. For assumption uncertainty, see AI mortality and longevity modelling.
A claim alleging that a report or advice letter was inaccurate or misleading would usually fall within actuarial PI, subject to the policy terms. Claims by third parties raise the separate question of whether a duty of care was owed to them, which commonly turns on the principles in Caparo v Dickman and on what the report said about who may rely on it. Clear reliance restrictions and agreed liability caps in your terms help shape exposure. See PI insurance for actuaries.
Apex arranges PI for actuaries and consultancies; a broker can check how your policy and engagement terms deal with third-party reliance on reports. See the AI in actuarial work overview.
Yes. No standard prohibits it, but you remain responsible for the communication under TAS 100 and Code Principle 6. Check every figure, make sure limitations and sensitivities are included, and use approved tools that keep client information confidential.
We have not found a specific rule requiring disclosure, but some clients may ask about AI use in their terms, so check your engagement letter. What matters most is that the content is accurate and complete and that you can stand behind every statement in it.
Possibly, if a duty of care is found on the facts. Courts commonly ask whether the actuary knew the third party would rely on the report for a particular purpose. Reliance restrictions and clear statements of purpose in the report help limit that exposure.
Tell us about your firm and the work you do. A broker will review how your current PI treats AI-assisted work, or arrange new cover. No fee for a quote. Or call 0117 325 0027.
Start a PI quote Request a callbackApex Insurance Brokers Limited is authorised and regulated by the Financial Conduct Authority. Registered in England and Wales, company number 07014570. This page is general information about how AI is used in the profession and how insurance may respond. It is not legal, regulatory or insurance advice on your circumstances, and it does not guarantee that cover will be available or on what terms.
