Companies are scrambling to rebrand themselves and their products as 'AI-focused' even when the underlying technology hasn't changed. The Guardian reports that this practice - dubbed 'AI washing' - is becoming so prevalent it's attracting regulatory scrutiny similar to 'greenwashing' in climate claims.

I've seen this movie before. During the dot-com boom, companies added '.com' to their names and watched their stock prices soar. During the big data era, everything was suddenly 'data-driven.' During the blockchain hype, companies with no crypto business pivoted to 'blockchain solutions.'

But AI washing might be worse, because AI is poorly defined enough that almost anything with an algorithm can claim to use it. That's not hyperbole. If your product uses any form of machine learning - even a basic regression model from the 1990s - you can technically call it AI-powered.

The pattern is predictable: companies facing stagnant growth or skeptical investors discover that adding 'AI' to their pitch decks increases valuations. Products that used to be 'automated' are now 'AI-driven.' Tools that used to do 'analytics' now do 'AI-powered insights.' It's the same functionality with better marketing.

What makes this frustrating is that it obscures actual AI innovation. There are companies doing genuinely novel things with large language models, computer vision, and reinforcement learning. When everything claims to be AI-powered, those real advances get lost in the noise.

Regulators are starting to notice. The Guardian reports that AI washing is attracting the same kind of scrutiny as greenwashing - making false or misleading claims about environmental impact. If a company claims AI capabilities it doesn't actually have, that's potentially securities fraud or false advertising.

The challenge is enforcement. How do you prove a company's 'AI-powered' claims are false when AI itself is vaguely defined? Is a rules-based system AI? Is basic machine learning AI? Where's the line between legitimate marketing and deceptive claims?

Here's my rule of thumb: if you can't explain what specific AI technique you're using and what problem it solves that traditional software couldn't, you're probably AI washing. 'Machine learning model' isn't specific enough. Neither is 'neural network.' Tell me the architecture. Tell me the training methodology. Tell me the actual innovation.

The technology is impressive when it's real. The question is whether we can distinguish real AI innovation from companies slapping 'AI-powered' on products that haven't fundamentally changed. Based on current trends, the answer appears to be no - not without regulatory intervention that's still years away.