Despite widespread use, trust in AI for fitness is low. In the United States, only 26% of people hold a positive view of artificial intelligence, according to an NBC poll, while in the United Kingdom 78% say they worry about negative outcomes from the technology. Yet usage tells a different story: 62% of Americans and 69% of Britons regularly employ AI in some form. This gap between adoption and confidence is sharply reflected in how people approach AI for exercise and diet – a tool many find useful, but one that leaves a significant portion of the public uneasy or outright hostile.
Readers who responded to a callout about AI and fitness illustrate the spectrum of opinion. Some rely on it to shape their workouts and meals. Others refuse to use it at all, citing concerns about the economy and the environment. Many fall somewhere in between – grateful for the convenience but wary of the broader consequences. The mixed sentiment echoes a wider cultural ambivalence: AI is increasingly embedded in daily life, but trust has not kept pace with uptake.
Personal fitness experiences: from lifelong routines to first-time swimmers
For some users, AI has been transformative. Jose, 60, from the Netherlands, turned to AI after repeatedly failing to maintain a steady fitness routine. He fed it details about his goals, past failures, a knee problem, and his former regimen as a high-school swimmer. The AI generated a weekly plan combining weight lifting at home – including exercises, loads and reps – with swimming three days a week. Two months in, Jose says he is sticking to a routine for the first time in years.
Dave, 62, a retired wine professional living in France, describes himself as having a “slovenly past” that led to weight gain. He credits his AI trainer with helping him shed the extra kilos. The AI set his indoor cycling program, schedule, dietary advice, provided feedback, corrected mistakes, and offered encouragement. “The AI had access to all the information needed online, and delivered it at the speed of light,” he says. “Trial and error was eliminated.”
Gerarda Cronin, 68, of British Columbia, learned to swim just a year ago. Now she swims at least three times a week using MySwimPro workouts on her Apple Watch, which are customised for her ability and goals. The AI sends upbeat coaching texts after each session, offering advice and celebrating achievements. She also swims with a Masters club where human coaches provide feedback on technique – something the app cannot do. “I think the two methods are complementary,” she says. “Today, I swam 2,800m. Not bad for an old gal who only learned to swim properly a year ago!”
Chris Morrell, 49, of Surrey, uses AI across most parts of his life, treating it as a personal trainer, personal assistant and doctor. He feeds it transcriptions of podcasts and other sources about longevity, and the AI develops plans for resistance training, rowing, and yoga. If he has an injury, he inputs that information to ensure the day’s exercises won’t aggravate it.
Tom Davison, 30, from the Netherlands, “vibe-coded” his own app so he and his girlfriend can work out together and hold each other accountable with features like streaks. The app suggests workouts based on their goals, sets a schedule, times reps or distance, and gives nutritional suggestions by tracking meals. “It took two evenings to build and is exactly tailored for me,” he says.
Debbie McGowan, 56, of West Lancashire, started taking Mounjaro in August 2024 and committed to strength training three times a week. She lacked the time and money for a personal trainer, so she turned to ChatGPT to analyse her progress and guide her. She continues to use it to develop programs, check form, and work around injuries – but also does her own research rather than rely solely on the AI. “It can be an excellent tool if used intelligently,” she says. “It can be wrong, and confidently so. It sometimes feels like working with an over-enthusiastic, knowledgeable, but inexperienced intern.”
Drew Long, 46, of Edinburgh, is not a fan of AI in general – he finds it overrated – but admits it is useful for straightforward tasks. After seeing the importance of strength training in online wellness spaces, he watched YouTube videos to learn form, then used ChatGPT to discuss his diet and get workout suggestions tailored to his lifestyle, age, and available time.
Patrick McCann, 62, of New Zealand, asked AI to devise a program to increase his bench press from 120kg to 140kg over 12 months after retiring from motorcycle racing. The AI incorporated some familiar and some unfamiliar exercises. He uses AI at work but says he feels “we were better off without them” – referring to AI, social media, the web, and most tech innovations since 1985.
Peter Saddon, 56, of Yeovil, wanted a personal trainer but was unimpressed by the one at his gym and found the cost prohibitive. He uses AI to build new gym programs, set weight targets and reps, and manage injuries and bad technique. “As long as you remember it is not an all-seeing god,” he says. “It reminds you from time to time how dumb it is.”
David Cooper, 61, of France, has a background in running and cycling. He uses AI to discuss his training sessions after uploading data, treating it like a one-on-one session with a coach. He does not rely on it for developing his plan – he has a human coach – but finds the combination positive. “I was interested to see how much AI would know about training principles, and was surprised at its knowledge and depth of understanding.”
Susan Mulhall, 76, of Washington DC, takes the opposite view. “I never use it,” she says. “AI and phones in the gym have been a source of irritation, isolation and argument. I loathe and despise AI.”
Expert caution: the risks of inaccurate advice and injury
The potential for AI to give incorrect fitness advice – and cause injury as a result – deserves the most careful attention. Dr. Carol Garber, professor of movement science and education at Columbia University, warns that developing a fitness program requires more than a “cookbook approach”, which is what AI typically provides. “AI samples from available information on the internet, and frankly, it often gets things wrong,” she says. Those mistakes could lead to serious injuries. Garber recommends working with a certified human exercise professional whenever possible, whether virtually or in person. If people do use AI for fitness guidance, she stresses the importance of always double-checking the information against credible sources.
This caution is particularly relevant in the UK, where professional bodies such as the Chartered Society of Physiotherapy and the Register of Exercise Professionals set standards for fitness practitioners. AI-generated advice may not align with those standards, and there are no specific UK regulations governing AI fitness tools. Data protection is another concern: the Information Commissioner’s Office oversees compliance with GDPR in the UK, and users of AI fitness apps should be aware of how their personal health data is stored and used. The environmental footprint of AI – the energy consumed by algorithms and data centres – also gives some readers pause, as reflected in the responses that cite the environment as a reason to avoid the technology.
The mix of enthusiasm and scepticism among users underscores the challenge. Many who rely on AI acknowledge its shortcomings. Debbie McGowan’s description of AI as an “over-enthusiastic, knowledgeable, but inexperienced intern” captures the tension. Peter Saddon’s reminder that AI is “not an all-seeing god” echoes the same note. Even David Cooper, surprised by AI’s depth of understanding, still prefers to keep a human coach in charge of his training plan. The technology offers speed, convenience, and personalisation, but it cannot replace the nuanced judgment of a trained professional who can observe form, understand biomechanics, and account for individual psychological and physical context. As Dr. Garber puts it, the risks of relying on a system that “often gets things wrong” are too great to ignore.
