01What happens.
The published record now answers investor questions through AI. Superseded results, legacy pages and guidance documents are read with equal confidence, and each repetition reinforces the wrong position.
For most people asking, AI's answer is the only version of the organisation they meet.
02Why it matters.
If you sit on the board or run investor relations, the disclosure record is now read by machine on behalf of the people you disclose to. It is read well beyond the current releases you check, into the releases you remember and the pages you do not. 79% of AI misinformation arises from organisations' own content. The first question an investor asks AI about the company is answered from that record, and the answer carries the company's name.
Regulators have told investors as much. The SEC, NASAA and FINRA issued a joint alert warning that AI answers can rest on outdated data & that investors should verify before acting.
03The case.
This illustrative scenario draws on patterns observed across listed company estates; no company is identified.
A listed company releases half-year results with revised guidance. The release goes to the regulatory news service, the investor site and the results webcast page. The full-year results from 2 years earlier, the capital markets day presentation and the trading update that preceded the revision stay where they were. They sit on the investor site, in the document store and on the financial news sites that republished them.
An investor asks AI what the company's guidance is.
The answer is the earlier range, stated plainly, without a date and drawn from the older documents, which are more numerous, more linked and better structured than the single new release. The investor's adviser asks a different tool and receives the same range, and an analyst's AI-drafted note carries it as well, because the same older documents are the ones each tool finds.
The investor relations director learns of the position from an investor, who asks why the company has walked back guidance it had raised.
The company has done no such thing. The explanation has to be made to each person who asks, and each explanation reads as a change of position, when the position changed months ago and the record did not move with it.
The 2-year-old results are meanwhile answering questions about the company's margins, the capital markets day deck is answering questions about strategy, and a page describing a business sold last year describes it as trading. Nobody in the company has published anything false, and nobody has looked at the record the way a machine reads it, because from the inside the current release is the one that counts.
04Exposure.
The board is accountable for the disclosure record, and the record now includes what AI says the company said. A superseded number in circulation as the company's current position draws the attention of investors and of the regulator alike, and the board's response has to begin by establishing what the company's position is. Whether any obligation is engaged is a question for the company's advisers on the facts; what is certain is that the board is answering for it.
The cost is the correction: investor relations time, adviser time, and the repeated explanation to people who acted on the wrong number.
The heavier cost is confidence in management, which goes when the market's version of the company and the board's version diverge. In industry benchmarking, 57% name AI errors and misinformation as the top concern threatening AI adoption (Gallagher 2026, respondents), and insurers are writing exclusions and endorsements on AI-related risk into the covers a board relies on. No outcome is automatic.
05What AI Readiness changes.
What the board cannot see from the inside is which parts of the record AI is answering from. AAAnow reads the company's digital landscape as AI reads it, evidences the superseded documents and legacy pages still answering as current, and places the position on the AI Readiness Maturity Scale™. With that evidence the board can decide, before the next results, which documents stand as the record and which are retired or marked superseded in a form a machine reads.
Your team makes the changes; AAAnow provides an assessment, the evidence, the guidance and the review.
06Related use cases.
The annual report AI read was not the one the board signed
The primary disclosure read wrongly because of how the file was built.
An old annual report answered as this year's after the page that dated it was removed
PDF authority outliving the page that framed it.
A hostile version of the results was the one AI quoted, because the company's own was locked away
The vacuum a company creates by blocking AI.
07References.
- SEC, NASAA and FINRA, joint investor alert on the use of artificial intelligence, warning that AI answers may rest on outdated data.
- 79% of AI misinformation arises from organisations' own content: AAAnow research.
- Gallagher, AI adoption and risk benchmarking 2026. www.ajg.com/news-and-insights/features/ai-adoption-and-risk-benchmarking-2026ajg.comWe do not enable external links - please copy and paste, you are then certain as to the link being opened
Next step
Talk it through, in your terms.
The briefing sets this case in your organisation’s terms, and the discussion follows your questions.
Illustrative scenario, drawn from observed patterns. No company is identified.