.png)
In short: AI washing means overstating how much artificial intelligence a product really contains. It has moved from a reputational problem to a legal one: the US SEC fined two investment advisers a combined USD 400,000 in 2024, and in 2025 prosecutors brought criminal charges against a founder whose "AI" app was largely run by people. The newest version is agent washing: Gartner estimates that only about 130 of the thousands of vendors claiming agentic AI offer the real thing. This article explains what AI washing looks like, why it is a leadership-level risk, and gives you a short checklist for testing any AI claim before you buy.
Artificial intelligence has become a fixture of business discussions. It shows up everywhere, from email filtering to logistics optimization and strategic decision support. There is a catch, though: not everything that is sold as AI actually is. For decision-makers, telling real value creation apart from empty promises has become an essential skill.
AI washing is when a company overstates or misrepresents the artificial intelligence in its products or services, either on purpose or because it does not fully understand the technology itself. The usual goal is a stronger market position or more investor confidence.
The term borrows from "greenwashing" in sustainability, where companies attach a fashionable, positive-sounding label to something that has little real substance behind it.
None of this is new. The same pattern played out during the "blockchain", "big data" and "IoT" waves. The pressure is real: a startup or scale-up that leaves AI out of its pitch can already find fundraising harder. But hype cycles end, and the companies that last are the ones creating measurable business value with technology that actually works.
AI washing is not only an ethical question. It now carries clear legal risk, and regulators on both sides of the business have started to act.
In March 2024 the US Securities and Exchange Commission (SEC) brought its first cases explicitly framed as AI washing and fined two investment advisers a combined USD 400,000:
Since then, the cases have become more serious. In January 2025 the SEC settled with Presto Automation, which had presented its drive-through ordering system as its own AI technology while relying on third-party technology and offshore staff to take many of the orders. In April 2025 the SEC and the US Department of Justice charged the founder of the shopping app Nate, which had raised more than USD 42 million on the promise of AI-automated purchases while contract workers placed most orders by hand. It was one of the first criminal AI washing cases.
Consumer and business marketing claims are under scrutiny too. The US Federal Trade Commission launched Operation AI Comply in September 2024, and according to Holland & Knight it brought more than a dozen cases linked to AI washing the following year, increasingly including B2B marketing claims.
As "agentic AI" became the buzzword of 2025, the same problem moved with it. Gartner calls it agent washing: relabeling existing chatbots, AI assistants and robotic process automation (RPA) tools as AI agents. Gartner estimates that of the thousands of vendors claiming agentic AI, only about 130 offer genuine agentic capabilities, and it expects more than 40% of agentic AI projects to be canceled by the end of 2027 because of cost, unclear value or weak risk controls.
This matters directly for procurement and supply chain teams. In May 2026 Gartner warned that vendors promising end-to-end autonomous supply chain planning before 2027 are overstating what is currently possible, and advised buyers to separate real capability from market noise.
AI washing creates three main risks for a company:
The team at Bobcats Coding, specialists in AI implementation and product development, built a simple framework for judging how much AI a product really contains. It helps leaders see where marketing ends and engineering begins.
Understanding different levels of AI integration in software
| Category | Characteristics | Example |
|---|---|---|
| AI-Native | AI is the foundation of the product, utilizing proprietary models and extensive datasets. | Autonomous vehicle control |
| Generative AI-Driven | The system is based on external models, such as OpenAI, but places a strong emphasis on the use of AI as a central value. | Specialized, industry-specific content generator |
| Generative AI-Assisted | AI serves as a supplementary feature; the product continues to function effectively without it. | Quiz evaluation module on an educational platform |
| Traditional software | AI is simply a convenience tool, such as predictive text input. | Modern email client |
AI is the foundation of the product, utilizing proprietary models and extensive datasets.
Autonomous vehicle control
The system is based on external models, such as OpenAI, but places a strong emphasis on the use of AI as a central value.
Specialized, industry-specific content generator
AI serves as a supplementary feature; the product continues to function effectively without it.
Quiz evaluation module on an educational platform
AI is simply a convenience tool, such as predictive text input.
Modern email client
Building on the Bobcats Coding research, five questions help you quickly test how credible an AI solution is:
There is a wide gap between companies riding the hype and those building real technology. Regulators have started to close that gap from their side. Leaders can close it from theirs by looking past the buzzwords and investing in solutions that deliver measurable efficiency gains and lasting competitive advantage, not just a good press release.
What is AI washing?
AI washing is when a company overstates or misrepresents the artificial intelligence in its products or services, usually to win customers or investors. It is the AI equivalent of greenwashing.
Is AI washing illegal?
It can be. In the US, the SEC fined two investment advisers a combined USD 400,000 in 2024 for false AI claims, and in 2025 the SEC and the Department of Justice brought civil and criminal charges against a founder whose "AI" app was largely run by people. The FTC also treats unsupported AI claims as deceptive marketing.
What is agent washing?
Agent washing is the newest form of AI washing: relabeling existing chatbots, assistants or RPA tools as autonomous AI agents. Gartner estimates that only about 130 of the thousands of vendors claiming agentic AI offer genuine agentic capabilities.
How can I tell whether a product really uses AI?
Ask the vendor to explain the model and architecture, where the data comes from and how it is maintained, and why AI is needed for the task at all. Then ask for a demonstration on your own data and for customers who use the feature in production.
.png)
.png)
.png)
.png)