Image credit: Getty Images
Image caption, Amazon has had to defend its use of AI technology in its physical stores Article information Author, Emma Woollacott Role, Technology reporter
4 hours ago
Amazon made headlines this year after reports questioned its “Just Walk Out” technology installed in many of its brick-and-mortar stores.
The AI-powered system will allow customers at Amazon Fresh and Amazon Go stores to simply select their items and walk out.
Only available to those who sign up through the app, the service uses AI facial recognition technology and an array of sensors and cameras to determine your selections, after which you'll be automatically charged.
Instead, the Indian workers were simply reviewing the system, Amazon said, adding that “this is no different to other AI systems that focus on accuracy and where human reviewers are the norm.”
However, Amazon also acknowledged that it would be reducing the number of stores that offer the Just Walk Out system.
Whatever the exact details of the Amazon case, it is a high-profile example of new and growing questions about whether companies are making exaggerated claims about their use of AI, a phenomenon known as “AIwashing,” after environmental “greenwashing.”
But first, let's remember what AI means exactly. Although there is no exact definition, AI enables computers to learn and solve problems. AI can do this after first being trained with vast amounts of information.
One particular type of AI that has been making waves in the past few years is so-called “generative AI” – AI that specializes in creating new content, such as through text conversations or composing music or images.
Chatbots such as ChatGPT, Google’s Gemini, and Microsoft’s Copilot are popular examples of generative AI.
There are several types of AI washing: some companies claim to be using AI when in fact they are using less advanced computing, others exaggerate the effectiveness of AI over existing technologies, or suggest that their AI solutions are fully functional when in fact they are not.
Meanwhile, other companies are simply adding AI chatbots to their existing non-AI operating software.
According to OpenOcean, a U.K.- and Finland-based investment fund for emerging technology companies, just 10% of tech startups mentioned the use of AI in their presentations in 2022, but that number is set to rise to more than a quarter by 2023. The firm expects that figure to exceed a third this year.
Sri Ayangar, a member of the OpenOcean team, also said that the competition for funding and the desire to be at the cutting edge tend to lead some companies to oversell their AI capabilities.
“Some founders seem to think that if they don't mention AI in their presentation, they're at a disadvantage, regardless of what role AI plays in their solution,” Ayyangar says.
“And our analysis shows a significant gap between companies that claim AI capabilities and those that demonstrate tangible AI-driven outcomes.”
Image courtesy of Sri Ayangar
Image captionSri Ayyangar says some startup executives feel they have to mention AI
It's a problem that's been quietly existing for years, according to data from another tech investment firm, MMC Ventures: A 2019 survey found that 40% of emerging tech companies that called themselves “AI startups” weren't actually using AI at all.
“The problem is still the same, but there's a different problem,” said Simon Menassy, general partner at MMC Ventures.
He explains that “cutting-edge AI capabilities” are now available to any company for the price of standard software, but rather than building an entire AI system, many companies are simply adding a chatbot interface on top of a non-AI product.
Dougal Dick, UK head of emerging technologies risk at accountancy giant KPMG, said the problem of AI-washing was exacerbated by the fact that there was no consistent definition of AI.
“If I asked people in a room to define AI, they would all give me a different answer,” he says. “The term AI is used very broadly and loosely, without clear standards. This vagueness allows AIwashing to emerge.”
“AI washing can have worrying implications for businesses, including overpaying for technology and services and failing to achieve the operational goals that AI was expected to help achieve.”
Meanwhile, investors may find it harder to identify truly innovative companies.
Additionally, Ayyangar said: “Unmet consumer expectations from products that claim to offer advanced AI-driven solutions could erode trust in startups that are doing truly groundbreaking work.”
Regulators in the United States, at least, are starting to take notice: Earlier this year, the Securities and Exchange Commission (SEC) announced charges against two investment advisory firms for making false and misleading statements about the scope of their AI uses.
“The tough stance taken by the SEC shows that there is no room for error when it comes to AI washing and, at least in the US, we can expect to see further fines and sanctions against those who violate its regulations,” said Nick White, a partner at international law firm Charles Russell Speechleys.
Image caption: Nick White says it's good to see US regulators cracking down on the issue
The UK already has rules and laws targeting AI washing, including the Advertising Standards Authority's (ASA) code of practice which states that marketing communications must not be materially misleading or likely to mislead.
Michael Corday, an associate in the regulatory team at UK corporate law firm Walker Morris, said AI claims were becoming increasingly common in ads that the ASA investigated.
One example is a paid Instagram post about an app with the caption “Enhance your photos with AI”, which the ASA found to be misleading and exaggerated the app's performance.
“What's clear is that AI-related claims are becoming more prevalent and perhaps more effective at generating consumer interest,” Corday said.
“In my opinion, we're at the peak of the AI hype cycle,” said Sandra Wachter, professor of technology and regulatory studies at the University of Oxford and a world authority on AI.
“But I feel like we forget to ask whether it always makes sense to use AI for every task. I remember seeing an advert for an AI-powered electric toothbrush on the London Underground. Who is this for? Who will it help?”
She also says that the environmental impact of AI is often underestimated.
“AI doesn't grow on trees… This technology already contributes more to climate change than the aviation industry. We need to move away from this one-sided, hyperbolic debate and think seriously about the specific tasks and areas where AI can be useful, and put it into practice instead of blindly deploying it in everything.”
But in the long term, the problem of AI washing may resolve itself, said Advika Jalan, head of research at MMC Ventures.
“AI, even if it's just a ChatGPT wrapper, is so pervasive that 'powered by AI' as a branding tool may not be a differentiator after a while,” she says. “It's a bit like saying, 'We're on the internet.'”
