01What happens.
Claims, product names and assets are published across regional sites, campaigns, partner material and historic records over many years. Heavey PDFs are only partly considered, AI fills in the gap s and then AI reads them together and blends them: it attributes claims to the wrong product, revives abandoned campaigns, repeats third-party wording as t he brand's own, and where the record runs out, invents. In a dispute, material still carrying the brand's name can be cited as evidence of its position.
02Why it matters.
The general counsel holds the register of what the company owns and may use: its marks, its licensed rights, the claims it can substantiate. The estate AI reads holds years of what the company once said, including marks it has sold, licences that have lapsed and claims it withdrew, and AI does not consult the register..
An AI answer that attributes an unapproved claim to the brand leaves the company having to disprove wording it never adopted for that product. AI governance needs an accurate view of the claims and assets still carrying the brand's name, alongside the legal register of what the company owns and may use.
03The case.
This illustrative scenario combines two documented patterns: AI systems fabricating product details and legal action over unlicensed use of protected assets. No individual consumer brand is identified in the scenario.
A consumer brand made a performance claim for one of its products at launch. Years (2+) later, after a complaint, the claim was withdrawn: the main product page was rewritten, the packaging changed, and the substantiation file closed.
The claim stayed where the withdrawal did not reach. A regional site, translated at launch and never re-translated, still carried it, and a campaign archive kept the launch pages live for reference. A retailer's product feed, built from the launch copy, still fed it to comparison sites.
AI, asked about the product, answered from the material that stated the claim in the most places and most clearly, and repeated the withdrawn claim as current. Asked about a sister product, it attached the same claim to that product, because the two shared a page template and the machine did not distinguish them.
Asked for evidence, it produced a figure from an unpublished draft in the repository.
A competitor's lawyers found the answers before the brand did. Their letter quoted the withdrawn claim, attributed to the brand by AI, as evidence that the brand still made it, and the unapproved figure from the draft as its substantiation. The legal team's register said the claim was withdrawn. The estate still carried it in numerous (5+) places, and an unpublished draft supplied a figure the brand had never approved.
04Exposure.
In this scenario, the competitor uses material still carrying the organisation's name to strengthen its case. The brand remains responsible for correcting claims on the regional site it controls and the campaign archive it kept. The unapproved figure from the draft creates a further dispute about a figure the brand never approved for publication. Marks and rights the brand no longer owns sit in the same estate, and the same mechanism can revive them as well.
Litigation cost is visible, while the brand also loses control over how its claims are presented. When a customer, retailer or lawyer consults AI first, the wider estate can shape their understanding before the legal register is seen.
05What AI Readiness changes.
The register recorded the decision, and the estate recorded the history, and only the estate was being looked at and consdiered. An accurate position of the landscape is the starting point: each claim, mark and asset still reachable under the brand's name, on the brand's own properties, on partner and retailer routes and in third-party repositories. That position shows the versions AI is drawing on and the claims it attributes to each product. We establish that position through the AI Accelerator and evidences it, then monitors it continuously, so that what AI says under the brand's name is overseen at scale.
The brand retires the material it controls and has partners and repositories correct what they hold.
The brand makes the changes, and AAAnow provides the assessment, the evidence, the guidance and the review.
06Related use cases.
Customers were shown a brand the company had retired, from guidelines it had replaced
The same weight-of-record mechanism applied to the brand itself, with marks no longer owned.
Drafts uploaded to the document store were served as the company's final word
The third-party repository that supplied the unapproved figure.
The impact of out-of-date safety information: who is liable?
Third-party reach defeating the brand's own record.
07References.
- AIMCLEAR, 2025: audit finding AI systems fabricating product details and features. https://www.aimclear.com/ai-search-fabrications-brands-ai-monitoring/aimclear.comWe do not enable external links - please copy and paste, you are then certain as to the link being opened
- Getty Images v Stability AI, 2025: legal action over unlicensed use of protected assets, as reported. https://www.theguardian.com/technology/2025/nov/04/getty-images-largely-loses-landmark-uk-lawsuit-over-ai-image-generatortheguardian.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 documented patterns. No company is identified.