KPMG withdraws AI report: hallucination and data reliability risk

The timeline: what happened with the KPMG report

On June 13, 2026, TechCrunch reported that KPMG has withdrawn a report on the use of artificial intelligence in organizations. The official reason: apparent hallucinations in the content generated by the models used in the research.

Therefore, the data, statistics, and claims present in the document were potentially unverifiable or directly incorrect. KPMG chose to withdraw the report from circulation rather than publish a corrected version. This decision, while courageous in terms of transparency, raised profound questions.

In fact, this is not a company new to AI. KPMG is one of the Big Four global consulting firms. However, even an organization with technical and human resources of that caliber produced a document riddled with generative errors. Consequently, the problem is not one of expertise, but of process.

Hallucination: A Working Definition for Those Who Don’t Work in AI

Hallucinations in language models are statements generated with apparent certainty but lacking any real basis. The model does not consciously lie. It simply produces statistically plausible outputs that may not correspond to the facts.

This phenomenon has been documented for years in technical literature. For example, MIT Technology Review had already conducted an in-depth analysis of the mechanism in 2023. Despite this, many organizations continue to use AI outputs without adequate verification protocols.

In particular, the risk increases when models are queried on quantitative data, industry statistics, or specific sources. Therefore, the KPMG case is not an anomaly: it is the predictable consequence of a content production process that does not include a structured fact-checking layer.

The winners and losers of this episode

Every instance of this kind leads to a redistribution of credibility in the market. Therefore, it is worth analyzing who gains and who loses ground in the wake of such an event.

Who loses: KPMG suffers direct, albeit temporary, reputational damage. Beyond this, the entire segment of research reports produced with generative AI loses credibility in the eyes of those who use them as a secondary source. Similarly, the AI model vendors involved—even if not explicitly named—see discussions about the reliability of their systems intensify.

Who benefits: Organizations that adopt hybrid approaches—AI supported by expert human review—emerge stronger from incidents like this. Furthermore, providers of AI governance and generative content audit solutions find a powerful sales pitch in this case. In summary, KPMG’s transparency in withdrawing the report, though costly, sets a positive precedent for AI crisis management.

Why Italian SMEs cannot ignore this signal

A common misconception is that incidents like this only affect large corporations with massive technology budgets. On the contrary, small and medium-sized enterprises (SMEs) are often more vulnerable because they have fewer resources for quality control of AI outputs.

Many medium-sized Italian companies now use AI tools to generate competitive analyses, industry reports, marketing content, and even business documents. However, these outputs are rarely subjected to systematic verification before distribution. As a result, the risk of publishing or sharing inaccurate information is real and often underestimated.

We of SHM Studio We observe this dynamic regularly in our work with clients. In particular, the problem emerges more frequently in sectors where market data changes rapidly: retail, manufacturing, and highly specialized B2B services. Therefore, AI content governance is not a theoretical topic; it's an immediate operational necessity.

SHM Studio Reading: A Process Architecture Problem

In our view, the KPMG case is not a failure of AI as a technology. It is a failure in the design of the process that led to the publication of the report. Therefore, the right question is not «Is AI reliable?» but «Is our production and verification process adequate?».

There are at least three levels of control that every organization should incorporate when using generative models to produce content with reputational or decision-making implications. First, verification of the primary sources cited by the model. Next, a layer of human review by experts with specific knowledge of the subject matter. Finally, a formal approval process prior to external distribution.

These three levels do not eliminate the risk, but they significantly reduce it. Furthermore, documenting this process protects the organization in the event of future disputes. The AI Consulting by SHM Studio includes the design of these operational workflows specifically for Italian B2B contexts.

Implications for those using AI in marketing and communication

Marketing is one of the areas where the adoption of generative AI tools has grown most rapidly in recent years. As a result, the implications of the KPMG case directly extend to the communications and content marketing functions of small and medium-sized enterprises.

For example, a company that uses AI to produce SEO copywriting, newsletters, or industry white papers must bear in mind that any automatically generated quantitative statement is potentially at risk. Similarly, competitive analyses produced using AI to support LinkedIn campaign o Google Ads campaigns require data verification before use.

Incidentally, the issue of the credibility of AI-generated content is also becoming a factor in organic search rankings. Google's guidelines increasingly emphasize firsthand experience and the credibility of sources. As a result, publishing content with inaccurate data generated by AI can harm both your reputation and organic visibility. For more information, see the section SEO by SHM Studio addresses these issues specifically in the Italian context.

What Nobody Tells You: AI That Talks About Itself

There is one aspect of the KPMG case that warrants separate consideration. The withdrawn report specifically concerned the use of AI in organizations. In other words, an AI model generated incorrect information about itself—or rather, about its own technological category.

This is not a minor detail. It indicates that language models face particular challenges when queried about up-to-date data, recent adoption statistics, or rapidly evolving industry benchmarks. In fact, training data always has a time cutoff. Therefore, any quantitative statement about recent phenomena—such as AI adoption in 2025 or 2026—is structurally at high risk of hallucination.

Research such as that conducted by McKinsey on the Global AI Survey show just how quickly adoption figures change. Therefore, using an AI model to cite statistics on these very trends is a high-risk endeavor without external verification.

Next moves: what to do now for organizations using AI

The KPMG episode suggests concrete actions for organizations that have already integrated AI tools into their processes. Below are the operational priorities that we at SHM Studio We recommend this to our B2B partners.

• Audit of existing processes: map where and how generative models are used, with a particular focus on externally distributed content.
• Introduction of a verification layer: Every quantitative or statistical claim produced by AI must be verified against primary sources before publication.
• Internal training: Teams that use AI must understand the mechanism of hallucinations and know when the risk is higher.
• Process documentation Formalizing the production and review workflow protects the organization and improves quality over time.
• Technology Partnership Review Evaluate whether the AI vendors used offer grounding tools, source citation, or integrated output verification.

Furthermore, it is worth considering that European AI regulations—particularly the AI Act—are introducing transparency and accountability requirements that will make these processes not only advisable but mandatory for certain categories of use. For this reason, investing in AI governance today is also an investment in future compliance. The section digital marketing and the one dedicated to web services at SHM Studio include consulting on these aspects for Italian SMEs. To learn more, you can Contact our team to explore other articles on the SHM Studio Blog.