Truace
Subscribe

AI-designed exercise plan helped older adults walk faster than standard balance training

Study title: Comparison of the effects of artificial intelligence, virtual reality, and traditional balance training programs on balance and gait in older adults

Researchers randomly assigned 48 retirement-home residents aged 60-75 to one of three eight-week programs: a virtual reality exergame, an exercise plan written by ChatGPT-4o and delivered by an instructor, or traditional balance training. Balance and gait measures were taken before, after, and at a two-month follow-up.

Read the study abstract

Purpose: With ageing, the ability to maintain balance and gait declines, leading to falls and reduced independence. This trial compared an AI-generated, human-delivered exercise program (ChatGPT-4o), a virtual reality (VR) exergaming program, and a traditional balance (TB) training program on balance and gait in older adults residing in retirement homes. Materials and methods: In this stratified, parallel-group, randomised controlled trial, 48 older adults aged 60-75 years were allocated 1:1:1 to VR (Xbox Kinect…

By the Truvace desk1 min read1 sourceEvidence: Moderate · 70/100 — Randomized controlled trial · n = 48 · 1 independent sourcePeer-reviewedTRV-2026-2331· cite & verify
Comparison of the effects of artificial intelligence, virtual reality, and traditional balance training programs on balance and gait in older adults
Hand grip strength and quality of life in older adults practicing Pilates method exercises title QS:P1476,en:"Hand grip strength and quality of life in older adults practicing Pilates method exercises " label QS:Len,"Hand grip strength and quality of life in older adults practicing Pilates method exercises " label QS:Lpt,"Força de preensão manual e qualidade de vida em pessoas idosas prticantes de exercícios do método Pilates: um estudo transversal" label QS:Lpt-br,"Força de preensão manual e qualidade de vida em pessoas idosas prticantes de exercícios do método Pilates: um estudo transversal" by Daniel Vicentini de Oliveira. CC BY 4.0 · https://creativecommons.org/licenses/by/4.0

In brief

Falls and loss of independence are major concerns with ageing, so cheap and scalable ways to design exercise programs matter. The findings suggest technology-assisted programs may add gait benefits, but the trial was small and limited to one age band and setting, and the AI only generated the program while a human delivered it.

Main points

  1. All three groups improved significantly over time on balance and gait measures.
  2. AI and VR groups gained more gait speed than traditional training, with no difference between AI and VR.
  3. VR produced the largest cadence gains, while AI produced the largest reduction in Gait Stability Ratio.
  4. Usability scored 82.4 for VR and 74.6 for the instructor-delivered AI program.
  5. No adverse events occurred during the trial.

The gain

An exercise program generated by ChatGPT-4o and delivered by an instructor produced greater gait-speed gains than traditional balance training and the largest reduction in Gait Stability Ratio among older adults.

The rundown

The trial was stratified and parallel-group, with 16 participants in each arm. Sessions lasted 60 minutes, three times per week for 8 weeks, followed by a 2-month follow-up.

Outcomes included balance, mobility, balance confidence, gait speed, cadence, Gait Stability Ratio, and usability on the System Usability Scale. Group-by-time interactions were significant for all outcomes except balance confidence (p = .063 after correction).

The authors conclude that AI-, VR-, and traditional training can each improve balance and gait, and that technology-assisted programs offer extra gait benefits supporting personalised, hybrid rehabilitation.

Sources

  1. Peer-reviewedDisability and Rehabilitation: Assistive Technology2026-10-31

The debate