[DigitalToday reporter Yoonseo Lee] A comparison has found ChatGPT presented a more structured workout plan than Siri AI.

On Aug. 1, IT outlet TechRadar reported that ChatGPT received a higher score in an expert assessment after both AI chatbots were asked for a workout routine under the same conditions.

The comparison was conducted by requesting a three-day split full-body workout plan that a 36-year-old man could do at home using only dumbbells in under 45 minutes. The goal was hypertrophy and muscle gain. With no follow-up questions, the test compared the services’ basic responses using only a prompt a typical user might enter in one go.

Siri AI set a routine with a push day focusing on chest, shoulders and triceps, a pull day focusing on back, biceps and rear deltoids, and a final schedule focused on lower body and core. It suggested 3 sets of each exercise, 8 to 12 reps, and 60 to 90 seconds of rest between sets, and instructed users to gradually increase dumbbell weight or the difficulty of bodyweight exercises. A strength was that it also briefly explained how to perform each movement.

But many days the actual time to complete the routine was often around 25 minutes, shorter than the suggested 45 minutes, and it also left out guidance on warm-ups and nutrition. Scott Laidler (스콧 레이들러), a personal trainer and fitness coach, assessed Siri AI by saying, "The structure is simple, so there is less room for misunderstanding," and rated its alternative exercises tailored to the environment as appropriate.

Laidler also saw Siri AI’s limitations as clear. He said Siri AI suggested increasing weight as fitness improves, but did not explain how to change the routine when its effectiveness declines after a certain point. He pointed out that the plan is designed for a limited period measured in months rather than years, and that it could lead to frustration unless users request a new plan periodically.

ChatGPT produced a longer-term plan. After first explaining workout priorities and basic principles, it presented a weekly schedule, warm-ups for each session, sets and reps for each exercise, and targeted muscle areas. It did not include movement descriptions, but added guidance on leaving 1 to 2 reps in reserve near the end, along with how to proceed, recovery, nutrition intake, and why this split is effective.

The expert assessment leaned toward ChatGPT. Laidler called the ChatGPT plan "a far more planned routine," saying Siri AI applied 3 sets and 8 to 12 reps across all exercises and stimulated each muscle group only once a week, while ChatGPT stimulated most muscle groups twice a week, varied sets by exercise, and included warm-ups and a progression rationale. He also said it was "more optimised for the stated goal of muscle growth."

Differences in additional features also emerged. ChatGPT could evaluate form based on a video uploaded by a user, but Siri AI could not provide such a feature.

There was also criticism that both services have limits as personalised workout coaches. Laidler said neither model asked follow-up questions needed to fine-tune the routine. If information such as injury history, perceived exertion and workout preferences is missing, a routine may not fit a user’s situation.

These limits also link to broader challenges for AI-based fitness advice. For users to get better answers, they need to include many details in prompts and request necessary follow-up questions in advance. At the same time, it remains that users need to know to some extent what a well-designed workout plan looks like. By current standards, ChatGPT is a step ahead of Siri AI in writing workout plans, but both services were assessed as unlikely to replace expert-level customised design.