Most of the fitness advice you have ever received was someone's opinion that got repeated until it sounded like a fact. "Eat six small meals to stoke your metabolism." "Lifting weights makes women bulky." "You have a 30-minute anabolic window after training." None of those survived contact with controlled research, and yet they are still printed in app onboarding flows and repeated by trainers today.
When I built Body by AI Coach, I made one rule that shaped everything else: a recommendation only ships if it traces back to the peer-reviewed literature. The methodology behind the platform is backed by 228 peer-reviewed citations. That number is not a marketing decoration. It is the reason the coaching gives different answers than the folklore you have been told your whole life.
What "Evidence-Based" Actually Means Here
Almost every fitness product on the market claims to be "science-backed." The phrase has been diluted into meaninglessness. For most apps it means someone read a blog post that summarized a study they never opened.
For Body by AI Coach it means something specific: the rules the engine uses to set your calorie targets, structure your progressive overload, sequence your training, and time your recovery are derived from the actual research record — randomized controlled trials, systematic reviews, and meta-analyses — not from the loudest voice in the comments section. When the evidence is strong, the engine is confident. When the evidence is mixed or thin, the engine says so instead of pretending certainty it does not have.
Why Folklore Is So Sticky
Bad fitness advice persists for the same reason any myth does: it is simple, it is repeated by people who seem confident, and it occasionally appears to work by coincidence. Someone eats six meals a day and loses weight — because they were in a calorie deficit, not because of meal frequency — and now they are a believer who tells everyone else.
The research record is the antidote to that anecdote-driven drift. A single person's experience cannot separate the variable that mattered from the variables that did not. A well-designed study can. The whole point of grounding the engine in 228 citations is that it is not susceptible to one loud person's coincidence.
Examples Where the Research Contradicts the Gym
Meal timing is the classic one. The popular belief is that frequent small meals raise your metabolic rate. Controlled research on energy balance does not support a meaningful metabolic advantage from meal frequency at a fixed calorie intake. The engine therefore does not nag you about eating every three hours — it focuses on total intake and protein sufficiency, which the evidence does support.
Soreness is another. Many people treat muscle soreness as the scoreboard for whether a workout "worked." The research on delayed-onset muscle soreness is clear that soreness is a poor proxy for the training stimulus that actually drives adaptation. The engine measures progress through performance and progressive overload, not through how wrecked you feel the next day.
Strength training and aging is the one I care about most. The folklore says older adults should "take it easy" and stick to light cardio. The research on resistance training in older populations says almost the opposite: progressive strength work is one of the most protective interventions available for maintaining independence with age. The engine does not soften your training because of your birth year — it calibrates it to your demonstrated capacity.
Honest About the Limits of the Evidence
Being evidence-based also means being honest about where the evidence is weak. There are populations the research has historically under-studied, and there are questions where the literature genuinely disagrees with itself. In those cases the right answer is not to fabricate confidence. It is to default to conservative, well-established principles and to be explicit that the recommendation is a careful best estimate rather than a settled finding.
That intellectual honesty is, paradoxically, the strongest signal that a system is actually built on research. Folklore is never uncertain. Folklore always sounds sure. The literature is full of appropriate hedges, and a coaching engine that respects the literature inherits those hedges instead of papering over them.
Why This Changes Your Results
If your coaching is downstream of folklore, your results are downstream of coincidence. You will chase meal timing and soreness and a hundred other variables that do not move the needle, and you will burn out before you ever get to the variables that do.
If your coaching is downstream of the research record, the engine spends your effort on the things that actually predict change: appropriate energy balance, sufficient protein, progressive overload, adequate recovery, and consistency over time. Twenty-eight citations short of two-hundred-thirty is a lot of reading. The point of doing it was so that you do not have to — and so that the advice you follow is the advice that holds up.