If you spend any time scrolling through social media platforms, you have likely encountered sleek and highly polished fitness videos promising dramatic physical transformations in a matter of weeks. These videos often showcase perfectly sculpted bodies, dramatic before-and-after comparisons, and bold claims that suggest you can look significantly younger or completely reshape your physique by following a simple routine.
At first glance, these results may seem incredibly appealing—but they often appear too good to be true. In reality, many of them are.
An investigation by the BBC has revealed that a growing number of fitness advertisements circulating online are misleading. These ads frequently feature characters generated by artificial intelligence rather than real human trainers or fitness enthusiasts. In many cases, the advertisers fail to clearly disclose that the individuals shown in the videos are not real people.
The purpose behind these deceptive campaigns is straightforward: to persuade viewers to purchase subscriptions to fitness applications or programs.
This raises important questions. How can viewers determine whether the person offering fitness advice actually exists? And perhaps more importantly—does it really matter?
Over the past few years, AI-generated content has rapidly taken over social media feeds. Fitness-related videos and advertisements have become particularly widespread, targeting users who are already interested in improving their health, appearance, or lifestyle. Once someone engages with this type of content, algorithms tend to amplify it, flooding their feeds with similar posts.
According to Professor Andy Miah, an expert in artificial intelligence from the University of Salford, this trend has grown significantly. He explains that people are naturally drawn to such content because they are actively seeking solutions for their fitness goals and personal appearance.
“There has always been demand for this kind of content,” he notes, “but now it has become extremely difficult to know who or what to trust.”
Unlike human influencers, AI-generated characters have the ability to produce an endless stream of content without limitations. Users cannot simply opt out of seeing this material, making it nearly impossible to avoid.
Professor Miah describes the current environment as a “wild west” in terms of regulation. While he acknowledges that artificial intelligence has many positive uses, he warns that misleading fitness claims can be harmful.
The issue lies in the unrealistic promises made in these advertisements. Claims such as losing massive amounts of weight in a short period or reversing the effects of aging are not supported by science. These exaggerated expectations can mislead viewers, giving them false hope and potentially damaging their mental well-being.
The investigation uncovered a variety of similar advertisements, all using different AI-generated characters but delivering nearly identical messages. Some featured a podcast-style interview where a supposed fitness expert claimed her routine could make women look decades younger within a month. Others depicted a fictional military instructor insisting that traditional gym workouts are ineffective, while promising extraordinary results through a specialized program.
In another example, three women were shown on a beach discussing their transformations, complete with before-and-after images. However, none of the individuals depicted were real. Similarly, a fabricated presenter claimed that medical professionals sought her advice, while promoting a routine that supposedly enables users to lose large amounts of weight in just a few weeks—all while being applauded by an artificial audience.
Fitness professionals in the real world strongly criticize these claims.
David Fairlamb, a veteran fitness instructor with over three decades of experience, expressed serious concerns about the impact of such misleading advertisements. Having worked in the industry long before the rise of social media and AI, he believes that while technology can be useful, it cannot replace genuine human coaching.
He emphasizes the importance of real interaction, accountability, and personalized guidance—elements that AI cannot fully replicate.
When shown examples of these AI-generated ads, his reaction was immediate and critical. He described them as misleading and potentially dangerous, especially for younger audiences who may be more impressionable.
According to Fairlamb, the idea of achieving a complete physical transformation in just 28 days is unrealistic. He insists that such results simply do not happen, regardless of the method used.
His daughter, Georgia Sybenga, who also works in fitness, agrees that even individuals familiar with social media can struggle to distinguish between real and artificial content. She admits that some of the AI-generated videos are so convincing that they can easily fool viewers.
Both father and daughter are concerned about the psychological effects of constant exposure to idealized and often artificial body images. They warn that young people, in particular, may develop unrealistic expectations about their own bodies, leading to dissatisfaction and mental health issues.
Another significant concern is safety. AI-generated fitness programs often fail to consider important factors such as injuries, medical conditions, or individual limitations. This lack of personalization could result in users following unsuitable routines and potentially harming themselves.
The Advertising Standards Authority has clarified that the use of AI in advertising is not inherently prohibited. Instead, the focus is on whether the content is misleading or harmful.
Adam Davison, the ASA’s director of data science, explained that the organization has received hundreds of complaints related to AI-generated advertisements in the past year, with the number continuing to grow. One of the key challenges is identifying whether AI has been used, as some content is highly sophisticated and difficult to detect.
He also noted that AI tools have made it easier for individuals and companies to produce advertisements quickly, sometimes without a full understanding of advertising regulations.
In response to the findings, the ASA has taken action against certain advertisers by issuing guidance notices. These notices are intended to educate rather than punish, helping advertisers understand their responsibilities and ensure compliance with advertising standards.
Social media platforms have their own policies regarding AI-generated content. They generally require that such content be clearly labeled. However, the investigation found multiple instances where disclaimers were either hidden, unclear, or entirely absent.
Major platforms like Meta and TikTok declined to comment directly on the findings. TikTok has stated that it has labeled over a billion AI-generated videos, while Meta uses detection systems based on indicators embedded in content creation tools.
Despite these efforts, many users have expressed a desire for greater control over the content they see, including the option to opt out of AI-generated material entirely. However, there is currently no clear indication that such features will be implemented.
As AI technology continues to evolve, the volume of generated content is expected to increase even further. Experts believe that the economics of social media—driven by attention and engagement—naturally favor the production of AI content.
While artificial intelligence offers many benefits, the potential for misuse remains a serious concern. When it leads to unrealistic expectations or misleading claims, there is a growing need for stronger regulation and greater awareness among users.
