While the benefits of flexibility exercises for office workers are well known, there is no randomized controlled comparison to determine whether these exercises can be delivered more effectively and safely by a specialist or by an automated artificial intelligence algorithm. Furthermore, considering the dynamic programming potential of artificial intelligence, determining whether this method can provide results similar to or superior to traditional methods is critical for the development of future healthcare models. The primary objective of this study is to compare the effects of 15-minute home-based flexibility exercises derived by artificial intelligence and produced by humans, administered over a 4-week period, on office workers' overall flexibility, back health, and muscle stiffness levels. This study is a two-armed, single-blind randomized controlled trial. Participants will be blinded to which group they are assigned to. This study will be respond "Are flexibility exercises generated by artificial intelligence more effective than human-generated flexibility exercises" question. During the study, participants who sign the volunteer information form and consent form and meet the inclusion criteria will be divided equally into two groups using a simple random method (coin toss):

• Group A: Artificial Intelligence (AI)-Based Flexibility Group.
• Group B: Human Expert-Based Flexibility Group. Intervention programs will include a standardized 15-minute flexibility exercise program will be automatically generated for each participant via Gemini, an AI-powered platform for group A, and a 15-minute program consisting of standardized flexibility exercises modified for each participant created by two qualified physiotherapists, based on calisthenic exercises (using body weight). This program will target the same muscle groups as the AI group's program and will be of similar duration. The exercise program will last a total of 4 weeks. Participants will exercise 3 days a week (Monday, Wednesday, Friday) for 15 minutes a day in their own environment.

This study will be a randomized controlled trial. The hypothesis of the study;

• HO1: Flexibility exercises generated by artificial intelligence improve overall flexibility in office workers more than human-generated flexibility exercises.
• HO2: Flexibility exercises generated by artificial intelligence reduce back pain levels in office workers more than human-generated flexibility exercises.
• HO3: Flexibility exercises generated by artificial intelligence are more effective than human-generated flexibility exercises in relaxing muscle stiffness in office workers.

Office workers employed at Erzurum Technical University will be included in the study. The minimum number of participants required for the study was determined to be 21 people per group using the Gpower 3.1 (Düsseldorf, Germany) program, with a 0.05 margin of error, 95% confidence interval, and 1.03 effect size.

All participants' pre-test (baseline) and post-test measurements at the end of the 4-week intervention period will be collected under the same conditions and in the same environment. The exercises for both groups will be provided to them as a brochure after the first measurement session, and instructions on how to perform them will be given. The intervention period will last for 4 weeks. On exercise days-Monday, Wednesday, and Friday-participants will receive reminder messages via the designated online communication group to remind them to do their exercises. Participants will be asked to provide brief feedback confirming they have completed the exercises for compliance tracking.

The AI-generated flexibility exercises will consist of a 2-minute warm-up, 11 minutes of stretching and flexibility exercises (including static and dynamic stretches), and a 2-minute cool-down. Human-generated flexibility exercises will consist of a 2-minute warm-up modified from calisthenic exercises, 11 minutes of main flexibility exercises (squats, modified push-ups, planks, etc.), and a 2-minute cool-down.

In our study, a demographic data collection form will be used to evaluate demographic data, the sit-and-reach test for flexibility assessment, the Oswestry Disability Index to assess back health, and the MyotonPRO muscle tone assessment device to evaluate muscle stiffness.

All measurements will be conducted in two sessions at the Therapeutic Exercise Laboratory of Erzurum Technical University. Measurements will be taken immediately before the start of the study and after the 4-week training period. An informed consent form will be obtained from each participant, and the test procedures will be explained in detail. Measurements will be performed by an experienced physical therapist in accordance with standardized protocols. Participants will be given sufficient rest time before each test, and their safety will be ensured. The total assessment time is estimated to be 12 minutes for each participant.

The obtained data will be analyzed using the SPSS 24.0 (Statistical Package for the Social Sciences) program. At the same time, comparisons will be made by entering the norm values into the SPSS 24.0 program. The Kolmogorov-Smirnov or Shapiro-Wilk tests will be used to examine whether the data are normally distributed. Descriptive statistics (mean, standard deviation) will be used for normally distributed data. The X2 test will be used for categorical variables. The independent samples t-test will be used to compare independent groups. Non-parametric tests (e.g., Mann-Whitney U test) will be preferred for data that do not show a normal distribution. For intra-group comparisons, the paired t-test or Wilcoxon signed rank test will be used depending on the normal distribution status of the data. In all statistical analyses, the significance level will be accepted as p<0.05.

Inclusion Criteria:

• Not have any known chronic orthopedic problems (active injuries requiring surgery or causing limited mobility).
• Be able to walk distances of 150 meters or more without assistance.
• Be an office worker who spends most of their working hours at a desk (at least 30 hours per week).

Exclusion Criteria:

• Having a chronic illness (cardiovascular disease, respiratory disease, metabolic disease, etc.).
• Having a condition that causes acute or chronic pain related to the musculoskeletal system.
• Having undergone major surgery within the last 6 months.
• Experiencing a health problem during participation that would cause any physical activity restrictions.