Amazon received critical headlines this year after reports questioned the “just walk out” technology installed at many of its physical grocery stores.
The AI-powered system enables customers to easily pick and drop off their items at its Amazon Fresh and Amazon Go stores.
Open to those who have registered through an app, AI uses facial recognition technology, and a range of sensors and cameras, to choose what you want. After that you will get the bill automatically.
However, it was widely reported in April that instead of using only AI, Just Walkout needs around 1,000 workers in India. About three-quarters of transactions are manually checked.
Amazon was quick to claim this. The reports were “false”. And that staff in India were not reviewing video footage from all stores.
Instead, it was said that the Indian workers were merely reviewing the system. Amazon added that it is “no different than any other AI system that places a high value on accuracy, where human reviewers are common”.
However, Amazon also confirmed that it will reduce the number of stores that use the Just Walk Out system.
Whatever the exact details of the Amazon case, it's a high-profile example of a new and growing question — whether companies are making too many claims about their use of AI. This is a phenomenon that has been dubbed “AI washing” in reference to environmental “greenwashing”.
But first, a reminder of what AI really means. Although there is no exact definition, AI allows computers to To learn and solve problems. AI is able to do this after first being trained on a lot of information.
The particular type of AI that has made all the headlines in the past few years is called “generative AI.” It's an AI that specializes in creating new content, whether it's text conversations, or producing music or images.
Chatbots such as ChatGPT, Google's Gemini, and Microsoft's Copilot are popular examples of creative AI.
There are many types when it comes to AI washing. Some companies claim to use AI when they are actually using less sophisticated computing, while others overstate the effectiveness of their AI over existing techniques, or suggest that their AI solutions Fully functioning when they are not.
Meanwhile, other firms are simply bolting AI chatbots onto their existing non-AI operating software.
While only 10% of tech startups in 2022 mentioned using AI in their pitches, this rose to more than a quarter in 2023, according to OpenOcean, an investment fund for new tech firms based in the UK and Finland. . He expects that number to rise by more than a third this year.
And, says OpenOcean team member Sri Iyengar, the competition for funding and the desire to appear cutting-edge have forced some companies to overstate their AI capabilities.
“Some founders believe that if they don't mention AI in their pitch, it can put them at a disadvantage, regardless of the role it plays in their solution,” says Mr. Iyengar. .
“And from our analysis, there is a significant disparity between companies that claim AI capabilities and companies that demonstrate tangible AI-powered results.”
According to data from another tech investment firm, MMC Ventures, it's a problem that has quietly existed for years. A 2019 study found that 40% of new tech firms that described themselves as “AI startups” actually were. Virtually no AI was used.
“Today the problem is the same, just a different problem,” says Simon Menashe, general partner at MMC Ventures.
He points out that “cutting-edge AI capabilities” are now available for every company to buy at the price of standard software. But rather than building an entire AI system, he says, many firms are simply layering a chatbot interface on top of a non-AI product.
Dougal Dick, UK head of emerging technology risk at accountancy firm PMG, says the problem of AI washing is not helped by the fact that there is no single agreed definition of AI.
“If I asked people what their definition of AI is, they would all give a different answer.” they say. “The term is used very broadly and loosely, without a clear point. It is the ambiguity that allows AI laundering to emerge.
“AI washing can have impacts on businesses, from overpaying for technology and services to failing to meet the operational objectives that AI was expected to help them achieve.”
Meanwhile, it can make it difficult for investors to identify truly innovative companies.
And, says Mr. Iyengar: “If consumers have high expectations from products that claim to offer innovative AI-powered solutions, it can erode trust in startups that are actually doing the work on the ground.” “
Regulators, at least in the US, are starting to take notice. Earlier this year, the US Securities and Exchange Commission (SEC) said it was Charging two investment advisory firms Along with making false and misleading statements about the extent of their use of AI.
“The firm stance taken by the SEC shows a lack of clarity when it comes to AI laundering,” says Nick White, “which suggests that, at least in the U.S., we Violators can expect further fines and sanctions.” , partner at the international law firm Charles Russell Speechless.
In the UK, there are already regulations covering AI washing, including the Advertising Standards Authority's (ASA) Code of Conduct, which states that marketing communications must not, or are likely to, materially mislead. .
AI claims subject to ASA investigation have become an increasingly common feature of advertising, says Michael Cordoux, partner in the regulatory team at UK corporate law firm Walker Morris.
Examples include a paid Instagram post about an app titled “Enhance your photos with AI”, which was commissioned by the ASA. exaggerated app performance, and was therefore misleading.
“What is clear is that AI claims are becoming increasingly popular and, most likely, effective in influencing consumer interest,” says Mr. Cordeaux.
“I think we're at the peak of the AI hype cycle,” says Sandra Wachter, professor of technology and regulation at Oxford University and a leading global expert on AI.
“However, I think we've forgotten to ask whether it always makes sense to use AI for all sorts of tasks. I remember seeing adverts on the London Tube for electric toothbrushes powered by AI. This What is it for? Who helps it?”
In addition, she says the environmental impact of AI is often understated.
“AI doesn't grow on trees… technology already contributes more to climate change than aviation. We need to move away from this overly one-sided debate, and really think about the specific tasks and sectors that So AI can be beneficial, and not just blindly implemented into everything.”
But in the long run, says Advika Jalan, head of research at MMC Ventures, the problem of AI washing may end on its own.
“AI is becoming so ubiquitous – even if it's just ChatGPT wrappers – that 'AI-powered' as a branding tool will no longer be a differentiator after a while,” she says. “It would be like saying 'we're on the Internet'.”