Risk · R-153 · Intellectual Property & Legal Liability
Overstated AI claims
The organisation misdescribes what its AI does, how well it performs or how it is governed, to customers, investors or regulators.
- Severity
- Medium
- Audience
- Board · Heads-of
- Frameworks
- PRIN 2A.5 · GOVERN 4 · Transparency
- Responsible AI domain
- Lifecycle stage
- Reviewed
- 2026-11-11
What it is
There is commercial pressure to say that a product uses AI, that the AI is accurate, and that it is responsibly governed. Statements run ahead of evidence: a rules engine is described as AI, a pilot as a deployed capability, an accuracy figure from a demonstration as typical performance, a policy document as a governance framework. The practice is often called AI washing. Statements to customers are subject to consumer protection and advertising law, statements to investors to securities law, and statements to regulators to the duty to be open with them.
How it happens
- Marketing copy is written without reference to test evidence.
- Claims made at launch are not revisited when the product changes.
- Responsible AI principles are published with nothing behind them.
- Sales teams improvise answers to questions about how the system works.
Why it matters
Regulators have taken action over misleading AI claims. Customers who relied on a claim may have a remedy. A published commitment on governance becomes the standard the organisation is judged against after an incident.
Warning signs
- Public claims about accuracy or capability cannot be traced to evidence.
- No one reviews AI statements before publication.
- Published principles have no owner and no reporting against them.
- Investor materials describe AI capabilities that are still in development.
Mitigating controls
Primary controls
- Review of AI-generated communications Preventive
- Reviews public and customer statements about AI against evidence. Customer communications produced by AI are checked against content standards before use, in full or by sampling according to risk, and required wording is enforced.
Supporting controls
- Independent assurance and internal audit of AI Detective
- Tests stated governance against practice. Second-line review and internal audit cover AI governance and the highest-tier systems on a risk-based plan, test systems as well as documents, and track their findings to closure.
- Model and system documentation Directive
- Provides the documented performance to check claims against. Each AI system's purpose, design, data, test results and limitations are documented to a set standard that allows a validator, auditor or regulator to understand it without its authors.
- Public AI transparency record Directive
- A public record of AI use disciplines what is claimed. The organisation publishes and maintains a record of its significant AI uses, describing what each is for, what data it uses and what safeguards apply.
Framework and clause references
| Framework | Clause | Title |
|---|---|---|
| FCA Consumer Duty (PRIN 2A) | PRIN 2A.5 | Consumer understanding outcome Communications must not mislead. |
| NIST AI Risk Management Framework (AI RMF 1.0) | GOVERN 4 | A culture that considers and communicates AI risk Honest communication of risk. |
| OECD AI Principles | Transparency | Transparency and explainability Transparency about AI systems. |
Sources
- OECD AI Principles OECD · official text
- NIST AI Risk Management Framework (AI RMF 1.0) NIST · framework
Related terms
How to cite
Axiom Verity (2026). Overstated AI claims. Reference Library, R-153. https://axiom-verity.com/risks/overstated-ai-claims. Reviewed 2026-11-11.
Found an error or a better source? Report a correction.