Wellness & Fitness
1 gain1 Trace0 problems
Every measured gain and problem on the record for this topic, newest first.Too few to weigh yet. Updated .
One square per record, darker for stronger evidence:Evidence: 0 high2 moderate0 emerging
2 records so far, too few to weigh yet. Is your study missing? Add it
0 records
No records of this kind in this topic yet.
AI in wellness and fitness
A Trace is one record with both a gain and a problem, so it counts on both sides: 2 records show a gain and 1 shows a problem. 2 context records are in the record list but not counted, because they report no measured gain or problem.
| Evidence | Gain side | Problem side |
|---|---|---|
| HighStrong study design, backed by at least two independent sources | 0 | 0 |
| ModerateA measured study or an official source | 2 | 1 |
| EmergingEarly, small, unreviewed or reported second-hand | 0 | 0 |
By type of evidence: 1 randomized trial; 1 benchmark or lab test. How evidence is scored
2 records so far, too few to weigh yet. A written summary appears once a topic has at least 20 counted records, with at least 5 on each side. Is your study missing? Add it
In a trial of 48 retirement-home residents aged 60-75, an exercise plan written by ChatGPT-4o and delivered by an instructor raised gait speed more than traditional balance training and cut Gait Stability Ratio the most.
In a test of 191 Spanish meal descriptions, all 20 AI model configurations gave more repeatable energy estimates for the same meal than the audited cascade comparator; the best reached a test-retest reliability (ICC) of 1.0000.
In a benchmark of 191 meal descriptions, the most repeatable model also gave the most incoherent answers: 65.58% of its responses had energy that did not match its own macronutrients under the 4/4/9 Atwater approximation, versus 3.80% for the audited cascade.
ace