Have you noticed that almost anything you want to buy these days is ‘AI enhanced’ or ‘harnesses the power’ of AI.
This isn’t just in the enterprise tech landscape either, you could even buy a type of CocaCola that was ‘co-created with human and artificial intelligence’... whatever that means.
In this article we’ll tell you everything you need to know about the AI hype cycle, AI washing, including the rise of agent washing, the cases regulators have already pursued, and how to spot it before you invest.
What is AI Washing?
AI Washing is the practice of exaggerating, or misrepresenting, how much artificial intelligence is really behind a product or service.
Law firm Orrick describes it as a company talking up its AI capabilities in order to mislead the market.
In practice, that can mean a simple rules-based system sold as AI, or one small AI feature used to brand a whole product as “AI-driven.”
The term hit the headlines in March 2024, when the US Securities and Exchange Commission (SEC) announced the first settled charges, against investment advisers Delphia and Global Predictions. The firms paid $400,000 in penalties over misleading claims about their use of AI.
What is Agent Washing?
Agent washing is the next phase of AI washing, especially across enterprise tech.
The term, defined by Gartner, demonstrates vendors rebranding existing products as agentic AI, such as AI assistants, robotic process automation (RPA) and chatbots, without any real agentic capabilities.
This confuses buyers by making it harder to tell which tools have genuine new agentic skills that are different to existing automation technologies.
It’s now a defining feature of the market. Gartner’s 2026 Hype Cycle for Agentic AI lists agent washing as one of its top three positions.
By Gartner’s estimate, only around 130 of the thousands of agentic AI vendors are the real thing.
Why Does AI Hype Lead to AI Washing?
Simply put, AI hype pays.
AI startups raised $220 billion in the first two months of 2026 alone, including a record $110 billion round for OpenAI. That’s more than 80% of the whole of 2025’s total, in just two months.
Forrester analyst Brandon Purcell told Axios that AI washing is tempting for startup CEOs because they know the funding will follow.
It’s not only investors feeling the pull, either. Nobody wants to be the business that missed the AI boat, so “AI-powered” becomes a box to tick rather than a claim to check.
Hype lowers our guard, too. When everyone expects AI to do everything, a product claiming to do anything sounds plausible.
When Vector RAG Stops Working
Why retrieval strategy now hinges on matching Vector, Graph and hybrid RAG to the questions AI must answer across complex enterprise data.
Is AI Hype Dying Down?
A 2026 study of 2,850 startups found AI and non-AI startups received almost identical funding, roughly $115 million versus $118 million. The research suggests that as AI becomes standard infrastructure, putting the label on a pitch deck doesn’t carry the weight it used to.
Meanwhile, Forrester says businesses are scrutinising the return on AI agents as the initial hype fades, warning that “every bubble inevitably bursts.”
When the hype deflates, the question changes from “does it have AI?” to “does it actually work?”. Gartner thinks the deflating has already started.
Is AI Washing Illegal?
There’s no specific law against AI washing, but that doesn’t mean it’s safe.
Misleading claims about AI are treated like any other misleading claim. That means they can fall foul of securities law, consumer protection law and, in the worst cases, fraud law.
After the SEC’s first cases, the FTC launched Operation AI Comply in September 2024, a crackdown on deceptive AI claims. Its chair said there is “no AI exemption from the laws on the books.” DoNotPay, which billed itself as a robot lawyer, agreed to pay $193,000.
In June 2024, prosecutors charged Joonko founder Ilit Raz with securities and wire fraud, alongside a parallel SEC case alleging she’d falsely claimed her recruitment platform ran on seven different AI algorithms.
The SEC’s enforcement director called it “an old school fraud using new school buzzwords.”
Inside the Agentic SOC Stack
See how unified telemetry, correlation engines and agentic AI workflows rebuild SOC architecture for autonomous detection and response.
How To Spot AI Washing/Agent Washing In Vendor Claims
It’s incredibly difficult to spot AI Washing in initial vendor claims, but most will not hold up to interrogation.
If a vendor can’t explain what the AI does, what data it runs on, or where it falls short, that’s your first red flag.
Be sure to ask how much human help the product needs in production. Ask for the real automation rate, not the demo version.
Be vary of agents that can’t act alone. Ask what the AI agentic system can plan and complete across several steps without a person or a fixed script. If the honest answer is “it’s a smarter chatbot” or “it’s RPA with a new logo”, you’ve found your agent washing.
Gartner says fully autonomous agents aren’t ready for most enterprise use cases, and that organisations should plan for semi-autonomous agents with human oversight. A vendor promising total autonomy should give you pause.
Similarly, be sure to ask for logs and observability, to test on your own data. Gartner calls AI observability one of the most significant developments on this year’s cycle, because it catches silent failures and outputs that sound convincing but are wrong.
If you can’t see what the AI is doing, you can’t account for it.
Comments ( 0 )