The Dangers of AI Washing

Fifteen years ago, when Big Data was all the rage, many companies that could barely manage small datasets branded their products as "Big Data solutions" to ride the wave of excitement around massive data handling and analytics. This created confusion in the market and made it difficult to distinguish between genuine innovations and those just using buzzwords for marketing. This phenomenon often leads to inflated expectations, making it harder for consumers to understand emerging technologies' value and capabilities. I see the same thing happening with Artificial Intelligence today.

What is AI Washing?

AI washing is when companies claim to use more AI in their products than they do. It's a marketing trick to make their products seem more advanced and take advantage of the growing interest in AI. "AI washing" is similar to "greenwashing," where companies exaggerate or lie about their positive environmental impact.

Some companies have been criticized for overstating the role of AI in their products to benefit from the trend. In March 2024, Security Exchange Commission issued its first civil penalties to two companies, Delphia Inc. and Global Predictions Inc., for making misleading claims about their use of AI.

Federal Trade Commission outlined the key questions they are asking companies to keep their AI claims in check:

1. Are you overstating your AI product's capabilities? Avoid making claims beyond AI's current limits, such as predicting human behavior. Deceptive claims lack scientific support or apply only under specific conditions.
2. Does your AI product outperform non-AI alternatives? You must have solid evidence to claim your AI product is better than a non-AI one. If such proof isn’t available, don’t make the claim.
3. Are you aware of the risks? Understand the potential risks of your AI product, such as failure or bias, before release. You can’t avoid responsibility by blaming third-party developers or calling the technology a “black box.”
4. Is your product really AI-enabled? Don’t falsely label your product as AI-powered. FTC experts can investigate whether your claims are valid, and using AI in development doesn’t necessarily mean the product is AI-driven.

So, what is wrong with AI Washing?

I have often joked in my customer presentation that we need to sprinkle AI fairy dust in our discussions because it almost feels like a right of passage in some business conversations. Several companies in the startup ecosystem are sprinkling this fairy dust frivolously to raise funds and increase their market caps.

The problem with AI washing is that it is outright deceptive. Imagine a toothbrush company claiming that its embedded AI will automatically clean your child's teeth without the child or the parent paying any attention to the brushing. Now, you buy the product and stop monitoring your child's brushing. After a year, the child's teeth are damaged. How are you feeling now?

AI washing can result in:

• Erosion of Trust: When companies exaggerate their AI capabilities, customers may be disappointed or dissatisfied when the product doesn’t deliver as promised, damaging the company's credibility and trustworthiness.
• Market Confusion: AI washing blurs the line between genuine AI-powered solutions and those merely using the term for marketing. This makes it harder for customers to identify true innovations.
• Hindered Progress: By overhyping basic tools as AI, companies divert attention from real advancements in the field, slowing down the adoption of genuine AI technologies.
• Regulatory Risks: As regulators increasingly scrutinize AI claims, companies engaged in AI washing risk facing legal penalties and damage to their reputation if their claims are proven false.

How to avoid AI Washing?

How can we avoid AI Washing, particularly when buying a new product? My one recommendation is to be skeptical of the marketing claims and always ask for evidence. Here are some red flags that you might want to consider:

• Lack of Specifics: Be wary if companies can’t explain which AI models, algorithms, or technologies they use, such as natural language processing or neural networks.
• No Transparency: Look for businesses that aren’t open about how their AI works, the data used, or the potential risks of their systems.
• Vague Claims: Avoid companies making broad, exaggerated statements like "AI-powered" without providing evidence or case studies to back up their claims.
• Failure to Address Bias: Ask how they handle data bias or AI hallucinations. If they can’t explain how they manage these problems, their AI may not be genuine.
• Absence of Case Studies or White Papers: Genuine AI use is often documented in detailed case studies or technical papers. If none are available, it may signal AI washing.

While AI can potentially transform industries, the market is filled with "snake oil" sellers trying to profit from exaggerated claims. It’s essential to be cautious and informed. In today’s world, AI literacy is no longer just a nice-to-have—it’s a critical business imperative for making smart, strategic decisions and avoiding being misled by empty promises.

If you like, please subscribe to the FAQ on Data newsletter and/or follow Fawad Qureshi on LinkedIn.