Athletifreak, a small athletic clothing company run by a New Jersey-based husband-and-wife duo, pitches itself online as a brand for active people born with “a freakish passion for movement.”
“Movement, to us, isn’t just physical,” Athletifreak, which has six brick-and-mortar stores primarily in malls in New York City suburbs, says on its website. “It’s a mindset. It’s how you show up when no one’s watching.”
Download the Straight Arrow app today to get the stories that matter free from manipulation, bias or agenda.™
Point phone camera here
There’s the chic woman relaxing on an urban stoop in an advertisement for the brand’s casual “Retreat” line. And there’s the ripped gym bro, repping Athletifreak’s “Weight Room” gym shorts while doing pullups on a playground.
Turns out, somebody was watching and, apparently, Athletifreak’s models have zero moves. That’s because they don’t even exist — not physically.
That is, at least, according to an anonymous complaint to the New York attorney general’s office, which accuses Athletifreak of failing to disclose its use of artificial intelligence-generated models to showcase its polo shirts, leggings and tank tops.
Athletifreak is among the first companies to face a consumer complaint of using “synthetic performers” — in place of humans — under New York’s first-in-the-nation law requiring advertisers to say when they use AI-generated models. Straight Arrow obtained the complaint through a public records request.
The law, which went into effect on June 9, comes at a moment when advertisements — and the internet more broadly — increasingly feature AI-generated images. The state synthetic disclosure law has been lauded as “a landmark first step” in AI transparency. Gov. Kathy Hochul, a Democrat, has championed the rules as a “simple, honest disclosure” that protects consumers from deceptive advertising and New York’s “creative workforce” from chatbots’ run for top modeling gigs.
Another anonymous complaint accuses British designer eyewear brand Bloobloom of targeting New York-based consumers with ads on social media “that feature completely artificial, AI-generated human models,” records show.
“The advertisements are distributed on digital and social media platforms [Meta] explicitly targeting consumers located in New York, routing traffic back to their e-commerce platform,” the complaint alleges.
Yet Athletifreak and Bloobloom are the only companies that have faced consumer complaints since New York’s law went into effect two months ago, a spokesperson for the attorney general’s office told Straight Arrow.
The spokesperson declined to comment on the status of the complaints. Neither Athletifreak nor Bloobloom responded to emails or LinkedIn messages seeking comment.
Perhaps one hurdle to enforcement? AI-generated images have become virtually indistinguishable from those that feature real people.
Spotting a fake
For Jacqueline Rodriguez, a 21-year-old who graduated from Duke University this spring with degrees in computer science and public policy, trying to spot AI slop on the internet has become a chore.
She said she uses social media only occasionally — a lot less often than other people her age, many of whom report being online “almost constantly.” But when she does visit TikTok or YouTube, she told Straight Arrow, the extent to which AI-generated images and text now dominate her feeds has become “very pervasive.”
Online communities are now devoted to spotting tell-tale clues like jumbled words and mangled fingers, but a body of research concludes humans are pretty bad at identifying fact from AI-generated fiction. The situation has left Rodriguez to constantly question whether an image she sees online is reflective of reality.
A year ago, “it was very obvious” when an image was AI-generated, she said, but the technology is changing quickly.
“There are times where it’s so realistic that, for a moment, you really see it as true,” she said. “It makes me really doubt everything that I look at and to make sure, ‘Is this real? Is this photo real?’”
In some cases, AI-generated images of real people, often called deepfakes, have been used for nefarious purposes, including for disinformation campaigns, harassment and to generate child sexual abuse material. Yet one recent study found that people are able to identify AI-generated content just 51% of the time — or, as the researchers put it, about “as good as a coin toss.”
Other reports have come to the same conclusion: That the human ability to detect AI-generated images is “only slightly above random guessing.” Such is true even with AI-generated images of people who are famous.
AI-generated images are now a fact of life in the advertising ecosystem, even as an overwhelming majority of people express hesitation and view disclosures as pivotal to trust.
Three-quarters of marketers have implemented AI into their workflows, according to a Salesforce report from February, and with great success. Chatbot-generated advertisements that “do not ‘look like AI’” have stronger click-through rates than those with images created by humans, according to a new Harvard University working paper.
Could AI disclosures improve trust?
AI-generated advertisements could be particularly effective in fashion, an industry that’s long used highly manicured, Photoshopped glamour shots to promote unrealistic beauty standards.
They’ve even made it to Vogue, though not without controversy. Social media users strongly criticized the magazine for publishing an ad by the clothing company Guess that featured two AI models.
Rodriguez likes fashion “quite a bit,” she told Straight Arrow, and brands can leverage AI to help her elevate her style. Eyewear retailer Warby Parker, for example, has a tool that allows people to see what they look like in different styles of glasses.
But when brands use AI-generated people to showcase their wares, Rodriguez said, they can be deceptive.
“They’re not a true display of a product and its usage or its appearance on an actual person,” she said. Disclosing AI use in advertising, Rodriguez said, could improve her trust that what she sees online reflects the reality of what she’ll get in the mail after clicking the checkout button.
“It just gives you knowledge and an informed understanding of what you’re looking at and not kind of feeling like you’re being tricked or something,” she said.
In March, Rodriguez published an article in the Duke Undergraduate Law Review that equated New York’s synthetic performer law to nutrition labels. The disclosure requirement, she wrote, could “reduce consumer deception and enhance informed choice.”
The Federal Trade Commission rules against “unfair or deceptive” advertising focus on preventing factual misrepresentations, she wrote, rather than “the authenticity of visual or audiovisual portrayals themselves.”
New York’s law closes that regulatory gap.
It is part of Hochul’s broader efforts to ensure “safe and responsible AI use,” a spokesperson for the governor’s office told Straight Arrow in a statement.
“During a time when technology is continuously evolving, is it important now more than ever to ensure the use of synthetic performers is clearly and honestly disclosed in order to protect consumers and respect those who work in creative industries,” the spokesperson said.
Why New York’s law affects all of us
Rodriguez lives in Dallas, Texas, and yet she said the New York law could ultimately influence what she — and everybody else — sees online. That’s because the rules apply to any business that markets its products to New York consumers. So, basically all of them.
Beyond the two consumer complaints with the New York attorney general’s office, the synthetic performer law has already led to changes at the globe’s largest online retailer, Amazon. In July, Amazon notified third-party sellers that any product images or videos featuring “AI-generated people” must feature disclosures that comply with New York’s law.
It’s unclear whether consumer complaints will lead to state enforcement actions, which include fines of up to $5,000 for repeat offenses.
The New York law, Rodriguez said, could soon prompt lawmakers in other states to act. If transparency becomes a regular part of the internet experience, she said, people could place greater value in content created by humans.
The rules could disproportionately affect small brands, like Athletifreak, that may lack budgets for studio photographers and Photoshop subscriptions, Rodriguez said. But it could ultimately lead to a better internet for consumers.
“If you see something that’s AI-generated, it’s not like it’s horrible or terrible,” she said. “But instead, there is more appreciation for when something is not AI-generated. You can actually see the work and creativity and authenticity of someone who isn’t just producing AI-generated content all the time.”
Round out your reading
- Former Fetterman staffers create social media account to expose former boss.
- Why Gen Z finally wants Boomer hand-me-downs.
- Want a copy of Pete Hegseth’s driver’s license? It’s for sale on the dark web.
- Smart glasses helped a man see the world. The privacy panic took it away.
- Americans are keeping cars longer, so the auto industry is making it easier to pay for repairs.
