AI washing—what it is and what the crackdown means for brands

Advertisers love talking about their use of AI, but doing so could put them in the sights of regulatory bodies cracking down on exaggerative or false claims regarding the technology. Last month, the U.S. Securities and Exchange Committee fined two organizations for “AI washing,” or falsely marketing their use of AI software. Two investment advisors were charged with making deceptive claims in materials familiar to advertisers, including press releases, social media posts and language on their respective websites. As a crackdown on AI washing emerges as a response to overhype, various players in the ad industry may be standing in its path, experts told Ad Age. Agencies and other firms developing ad tech, for example, are wont to dub new products as “AI-powered,” despite many of these tools—potentially three out of every four—having nothing to do with AI at all. The term “ChatGPT-wrapper” has emerged to pejoratively refer to this practice, said Gabriel de la Rosa Cols, principal at PR agency Intelligent Relations. It describes a product, service or entity that implies having a robust AI framework, but in reality is a basic custom tool built on top of another company’s AI platform, such as ChatGPT. “It’s like saying you’re a tech company because you use computers,” de la Rosa Cols said. Check out the latest Ad Age events and award programs Increased regulatory vigilance of AI comes as the technology is beginning to show growing pains. Platforms are acting up, driving fear into the minds of potential enterprise clients and spurring some brands to speak out on their shortcomings. Production companies are also grappling with the tension between the promoted benefits of AI and the real threats that it poses to creatives. The likelihood is significant that regulatory oversight of AI washing will eventually expand to business spaces including advertising, said Toby Coulthard, chief product officer of martech company Phrasee. The Federal Trade Commission hinted at this progression last year, when it issued a warning to advertisers to keep their AI claims in check. Snuffing out this kind of misleading information, wrote the FTC attorney who authored the admonition, is its “bread and butter.” Read more news about AI from Ad Age Start-ups across verticals, including advertising, are also at high risk of AI washing, de la Rosa Cols said, because they are incentivized to promote an AI narrative or else risk losing out on much-needed funding. In August, the FTC shut down an e-commerce startup called Automators AI for making false claims that it was using “AI machine learning” to maximize revenues, among other deceptive practices. Coulthard said he has seen a number of martech companies recently rebrand years-old technology around AI. Send-time optimization tools, for instance, which previously were not powered by AI are utilizing the tech, such as in Movable Ink’s Da Vinci AI engine, he said. Movable Ink has used machine learning to power its send-time tools ever since launching them, the company told Ad Age. Public, consumer-facing brands “should be the most cautious” when it comes to AI washing since regulatory bodies such as the FTC are more concerned with instances of consumer harm, said Katie Gardner, a partner at law firm Gunderson Dettmer. This does not mean, however, that enterprise-focused firms are off the hook. Both business-to-business and business-to-consumer companies face an array of legal repercussions for AI washing, Gardner said. Fines are most likely, as in the case of the two investment firms busted by the SEC last month, which were forced to surrender a combined $400,000. Direct consequences may also include lifetime bans from selling certain kinds of products and ongoing, mandatory audits. How to avoid AI washing The most important step a company can take when marketing its AI usage is to be specific, both about the kind of technology and the way it’s being used, said Phrasee’s Coulthard. Machine learning, chatbots, neural networks and large language models are just a few of the types of technologies that can reasonably be referred to as AI, and their use cases are virtually infinite. Since regulatory bodies are concerned about deceptive marketing, being as clear as possible is a way to mitigate false promotion. “No one has really defined what AI is … which gives companies a lot of latitude to claim it,” Coulthard said. Coulthard also suggested that companies be specific about where they’re not using AI, especially when the bulk of a tool or system’s capabilities belong to this category. Not only could this demonstrate to customers that the company is not seeking to inflate the value of its products, but may also suit the needs of clients—more of whom are requesting that agency partners do not use AI for certain tasks, Coulthard said. Companies should consider applying to AI the playbook on avoiding greenwashing, said de la Rosa Cols, because this literature is clear and well-documented. For example, the FTC cautions against brands calling themselves “eco-friendly” because the term is too vague and can confuse customers. An AI version of this guideline could discourage making equally vague claims about one’s AI use, such as “AI-powered.” As a practice of double-checking, companies may want to include implementing a new layer of internal review and approval, said Gunderson Dettmer’s Gardner. This means establishing a consistent communications channel with marketing and legal teams so that they can understand and evaluate the claims about AI being made. The information that these teams will want to check, said Gardner, includes which technology is proprietary and which is developed on top of a third party’s platform, such as ChatGPT. And companies that create products on top of third-party AI tools must realize they could be on the hook for that technology’s biases or errors, Gardner said. These platforms continue to show their many shortcomings, and while a PR disaster is one consequence, regulatory bodies will be watching to see if other consequences are merited. “We’ll see a lot more enforcement activity this year,” said Gardner. AI marketing blog How brands are using ChatGPT, DALL-E 2 and other tools Read more here

AI washing—what it is and what the crackdown means for brands

Featured Stories