AI-washing: What it is, and how to spot it?

A person holding out their hand with a digital AI symbol.
(Image credit: Shutterstock / LookerStudio)

A great deal of innovation has happened since AI exploded into the mainstream with the advent of ChatGPT. However, the AI boom has not come without its own inopportune side effects. One such side effect has been companies selling false promises and vaporware in the name of AI—a trend called “AI-washing”.

AI-washing can be explained by looking at its predecessor, “greenwashing”. This is when an organization gives misleading information about whether its product, services, supply chain or practices are ethically and environmentally sound. Greenwashing has become a known red flag to look out for, among consumers and enterprises alike. AI-washing, by comparison, is far newer and so is currently harder to spot. But there are a few things buyers can look out for along their AI investment journey.

The AI mislabeling trap

The term AI covers a pretty broad spectrum, which is partly why it’s so easy to make exaggerated promises about it. Generative AI, for example, uses machine learning to generate content based on prompts, all in real time. A generative AI product capable of summarizing complex conversations between multiple people, for instance, requires a significant investment in Large Language Models, a team of AI experts to train those models, and the ability to roll it out at scale.

By comparison, what many companies are currently selling is narrow AI—such as a phone’s virtual assistant—designed to complete very specific and limited tasks. While it can improve efficiency, it doesn’t have nearly as much power as the kind of generative AI that powers popular applications like ChatGPT. The reality is that few companies in the world are actually investing in generative AI. But plenty of companies will charge high prices for products that claim to be built on AI proprietary technology when they’re actually based on third-party AI technology or are still in their conceptual phase. Companies that own their own AI stack have greater control over data, privacy, security, performance and cost - key topics of discussion by organizations and regulators currently.

Brian Peterson

Co-founder and the Chief Technology Officer at Dialpad.

Asking the right questions

Companies can take a few approaches to ensure what they’re buying is real, provides value, and is the right fit for the business.

Is it available now? The term “vaporware” refers to software that’s being advertised but isn’t generally available yet—this can sometimes be hard to identify. Asking a direct question on the product’s availability will confirm whether you’re buying an actual working product, or simply joining a queue for something still in development.

- Is there an extra fee? It’s common for companies to advertise the availability of new AI features, but failing to mention the added cost per seat or per event. AI is expensive to run, and most companies use a third party to do so, thereby adding extra costs. Companies looking for new AI tools should focus on partnering with providers that build their own AI technology to get more bang for their buck—including customizable options, and greater data privacy.

- Can I get evidence of ROI? AI products are expensive, and so seeking an ROI report or calculator is crucial. Particularly when numbers can be inflated when mixed in with information about the main product, it becomes hard to understand the value of the AI versus the ROI on the product itself.

- Can I schedule a live demo? When it comes to investing in AI capabilities, a presentation isn’t going to cut it. Buyers shouldn’t settle for anything less than a live product demo before making any commitments—it should be considered a deal breaker.

Finding the best AI for your business

Once it’s been established that a product actually does what it says on the tin, buyers still need to pick the right kind of AI to support their business goals. Luckily, this is pretty simple when you track back to the core fundamentals.

Firstly, they must ask the question: is this a good product regardless of AI? In many cases, people will buy traditional software that’s AI-enhanced, rather than purchasing an AI-native product. It’s important to step back and consider why a product is being bought in the first place. For instance, when buying a CRM, if the quality of core functionalities is lacking but the AI seems good, it’s likely not worth it because the foundations aren’t there. The AI won’t necessarily mean the CRM is better.

The second question to ask is: do I believe this vendor will continue to innovate? Investing in an AI product is also a vote to future-proof a business. As such, it’s not just about what the product does now, but how it will evolve—for instance, will upgrades be made available without requiring the buyer to migrate its data? Choosing a vendor with proven track records in the AI space is like applying a safety net to a tech stack, and, therefore the overall business.

AI has the potential to improve efficiencies, ROI, and so much more for any business making the investment. However, it’s crucial for businesses to educate themselves and take steps to protect themselves against AI-washing. With the right AI, businesses can work faster, require fewer resources, and open up their workforce’s collective brain power on innovative tasks. When sourced correctly, AI can truly be transformative to any organization.

We've listed the best AI chatbot for business.

This article was produced as part of TechRadarPro's Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here:

Brian Peterson is a Co-founder and the Chief Technology Officer at Dialpad. Previously, he was a Senior Software Engineer at Google, building the front end of Google Voice.