AI in Strength Training: Programming, Feedback and Human Coaches

Artificial intelligence (AI) has rapidly penetrated deep into the fitness industry. It is promising to revolutionize how coaches design training programs and how athletes train. From smartphone apps that generate workouts to wearable sensors that track every repetition (reps), AI technologies are now assisting many tasks traditionally performed by human strength coaches.

The appeal is clear: AI can process huge amounts of data in seconds, be available 24/7, and offer personalized guidance at a fraction of the cost of a personal trainer.

AI-driven fitness apps raise an intriguing question: Can these machine-based coaches actually outperform or “outcoach” human coaches in strength training?

AI-generated strength training programs

A primary duty of a strength coach is to develop a personalized training program tailored to an individual’s goal. AI has shown considerable potential in this area by leveraging data and algorithms to generate workout plans.

For example, machine learning algorithms can analyze a user’s profile and past workouts to prescribe sets, reps, and weights that increase progressively over time.

FitnessAI’s developer describes adaptive workout programming and automatic changes to weights and repetitions as users progress. These are product claims, not independent evidence of optimal programs or a demonstrated advantage over a human coach.

Personalization at scale

AI’s ability to analyze each individual’s data and goals to create a custom plan instantly is something a single human coach would struggle with for thousands of clients. The quality and individualization of generated plans can vary, as the cited GPT-4 evaluation illustrates.

AI tools can assist with workout programming. Their availability does not establish what proportion of personal trainers use them.

Algorithms can use fitness data, performance records and goals to adjust routines. In Ghezelseflou and Choori’s interview study, athletes described perceived personalization and efficiency; the study did not test optimal loads or measured detection of overtraining.

Limitations of AI workout planning

However, current AI still has notable limitations in program development when compared to experienced human coaches.

In Dergaa and colleagues’ 2024 paper, experts evaluated 30-day GPT-4 exercise programs for five hypothetical profiles with different health conditions and goals. They found general safety-conscious plans but limited precision and individualization. This evaluated generated plans, not outcomes in patients who followed them.

The evaluation highlighted limitations in condition-specific programming and the need for expert oversight; it did not measure treatment outcomes or prove superiority over every human coach.

In Cori Ritchey’s 2024 first-person account, a later request produced a different plan after she supplied the same substantive information with different wording. She found that the plan met her requested parameters without necessarily being optimal. This experience does not establish results for identical prompts or identify one causal model mechanism.

Wearables and exercise tracking

The observation of real-time exercise performance by strength coaches represents another essential opportunity for them to provide feedback. This market is significantly influenced by wearable technology devices.

The Apple Watch contains built-in accelerometers together with gyroscopes which track movements. Businesses train their AI models to identify particular exercises through analysis of sensor data while simultaneously measuring athletic performance.

Train Fitness, now branded Motra on the developer’s website, advertises automatic exercise detection and repetition tracking from smartwatch motion. These are manufacturer-described features, rather than independent proof of coach-equivalent monitoring.

Some newer designs include systems capable of identifying breakdowns in exercise form and fatigue conditions. The combination of sensors and artificial intelligence systems has shown to be effective for monitoring exercise technique quality according to sports technology research. Weight machines equipped with sensors led to machine learning to evaluate leg press form that matched trainer evaluations according to a 2013 study. The instrumented-machine research should not be treated as validation of form-warning accuracy in every smartwatch or wearable. Such feedback may help users notice movement patterns, but it does not by itself establish injury prevention.

Computer vision and real-time feedback

Computer Vision (CV) is another powerful tool bringing real-time feedback to those training without a live coach present. Camera-based AI systems can visually monitor workouts and provide instant feedback critiques.

A high-profile example is Peloton Guide, a device introduced by Peloton that uses a smart camera connected to a TV to track exercises at home (Plumb, 2022).

Peloton’s 2022 launch described movement-completion tracking and Self Mode for visually comparing the user with the instructor. The Movement Tracker did not count repetitions or correct exercise form.

Completion and visual-comparison cues illustrate a different kind of feedback from a live coach. The 2022 Movement Tracker should not be described as giving repetition counts such as “only 5 push-ups left.”

What movement data can miss

Despite these advantages, limitations remain before AI can fully replace a coach’s real-time guidance. Current AI feedback is largely based on what it can measure and quantify.

A human coach’s feedback often goes beyond that, catching subtle cues like an athlete’s facial expression, mental state, or adjusting instructions based on personality. AI systems are not yet attuned to such nuanced observations.

They also lack the physical intervention capability. Additionally, setting up cameras or sensors for AI analysis can be a barrier.

The cited 2013 research illustrates a specific use of AI for evaluating instrumented weight-training movements; it does not establish general superiority over human coaches. The author discusses these tools as possible aids to coaching, rather than presenting evidence of how widely coaches have adopted them.

Motivation and automated coaching

A good coach does more than dictate workouts. They also motivate their clients and keep them consistent. Building strength requires motivation to adhere to a rigorous training regimen. Many fitness apps send push notifications reminding users to complete workouts and congratulate them on achievements. Chatbot-based AI coaches also attempt to engage users in a more interactive way to influence motivation.

When WHOOP introduced WHOOP Coach in September 2023, its announcement described GPT-4-powered conversational responses using the member’s data. Such feedback may offer a sense of availability, but the announcement does not measure motivation or establish a current model specification for every later version.

Additionally, some AI coaches attempt motivational interviewing techniques or positive reinforcement in their text-based interactions.

In Ellis and colleagues’ 2013 single-group, nonrandomized feasibility study, twenty community-dwelling people with Parkinson disease used a virtual exercise coach for one month. Mean adherence to daily walking was 85%.

Participants in that study not only stuck with the program but also reported high satisfaction with the virtual coach, rating their experience ~5.6 out of 7 on average (Ellis et al., 2013)​. This suggests that an automated coach can successfully encourage consistent exercise in the short term.

The human side of motivation

However, when it comes to human motivation, AI often lacks the deeper psychological influences that human coaches exert. Motivation is not purely about reminders or data, it’s also emotional and interpersonal.

Human coaches excel at providing social support, encouragement, and accountability in a way machines struggle to replicate. For many athletes, knowing that a person cares about their progress may be a powerful motivator.

The author sees accountability to a real person as one reason athletes may value human coaching.

That coaching preference does not establish one motivational response to every AI app.

Ghezelseflou and Choori’s 2023 qualitative study interviewed eighteen athletes. Participants described perceived personalization and training efficiency, alongside concerns about technical issues, privacy and personal touch. These are interview perceptions, not measured proof of improved athletic performance.

Trust and the coach–athlete relationship

The core of strength training coaching is the real bond developed between the coach and the athlete. A strong coach–athlete partnership depends on trust combined with effective communication and emotional backing. Athletes seek help from their coaches for both training lifts and developing healthy mental strategies. An uncertainty regarding AI strength training coaching lies in its inability to replicate human connection between coach and athlete. AI systems today function without true emotional capabilities because they rely solely upon processed information to formulate their responses. A human coach can demonstrate natural empathy towards athletes which enables them to create bonds that provide both motivational support and emotional comfort.

The author argues that analyzing data and offering feedback do not replace the personal relationship between an athlete and a coach.

Beyond providing instructions coaches serve as motivators who become trusted friends. A good coach understands emotional intelligence through which they create specialized communication methods and support structures for each athlete based on their emotional character.

Emotional connection and AI coaching tools

Trust is another factor discussed in athletes’ perceptions of coaching relationships. AI-generated results can vary with the system, inputs and context; the author argues that they do not replace a human coach’s personal understanding of an athlete.

The use of AI coaching tools in sports resulted in initial doubts among athletes. They wanted AI systems to display human characteristics that include empathy and personality while creating a feeling of understanding (Ghezelseflou & Choori, 2023).

Some organizations have developed chatbot coaching systems that ask users about daily activities or provide post-workout recognition. The author questions whether automated praise carries the same personal meaning as encouragement from a trusted coach.

Hybrid coaching: AI with human supervision

Not every athlete requires or desires a profound emotional bond. The basic requirement for some athletes includes both intelligent programming and technical support that AI systems can execute proficiently.

Hybrid models have become more popular because they unite AI-generated plans with human supervision. A hybrid approach can combine AI-generated suggestions with human review and support; this is a coaching option rather than a verified requirement of every platform.

The main strength of AI lies in its ability to process numerical data while making optimized plans.

Athletes continue to rely on human mentors for emotional encouragement and confidence development while seeking guidance from them. The coach handles emotional guidance while AI runs spreadsheet operations.

How AI can assist strength coaches

Technological devices can provide plans, monitoring and encouragement messages, as illustrated by WHOOP’s 2023 product announcement. Product availability alone does not establish improved training outcomes.

Coaching beyond measurable statistics

The advanced nature of these AI systems does not capture the entire essential aspects of coaching.

Real coaches provide intangible yet essential elements which cannot be readily quantified. These include genuine motivational support and the skill to understand athletes’ true performance status beyond the cold statistics.

Coaches adjust by watching their athletes’ tone of voice and body movements and by relying upon their instincts.

The author views emotional support as an important part of coaching, especially when training becomes challenging. Automated encouragement should not be treated as proof that a system understands an athlete’s experience.

Keeping the human coach in the loop

AI demonstrates its most effective use by assisting human coaching activities rather than serving as a replacement for human coaches.

Athletes often integrate AI-generated plans into their routines by partnering with real coaches for guidance through feedback sessions and motivational support.

Coaches can work together with AI systems to manage data processing and efficiency while humans maintain their role in emotional support (Dergaa et al., 2024). AI tools can support several coaching tasks, while the author argues for human oversight and relationship support. That trust and real connections are something athletes still need.

AI has made astonishing advancements to the fitness and personal training world. But ultimately, it needs fine tuning before surpassing in-person coaches or enabling coaches to perform their tasks at an elevated level.

References

360 Research Reports. (2021). Digital fitness market size, share and industry forecast 2021-2026. https://www.360researchreports.com

Alumni Ventures. (2023). Train Fitness is revolutionizing AI-powered strength tracking. https://www.alumniventures.com

Athletica.ai. (2024). Hybrid coaching: Combining AI precision with human intuition. https://www.athletica.ai

Dergaa, I., et al. (2024). Using artificial intelligence for exercise prescription in personalised health promotion: A critical evaluation of OpenAI’s GPT-4 model. Biology of Sport, 41(2), 221–241. Published online December 13, 2023. https://doi.org/10.5114/biolsport.2024.133661

Ellis, T., Latham, N. K., DeAngelis, T. R., Thomas, C. A., Saint-Hilaire, M., & Bickmore, T. W. (2013). Feasibility of a virtual exercise coach to promote walking in community-dwelling persons with Parkinson disease. American Journal of Physical Medicine & Rehabilitation, 92(6), 472–481. https://doi.org/10.1097/PHM.0b013e31828cd466

Ghezelseflou, H., & Choori, A. (2023). Athlete Perspectives on AI-Driven Coaching Technologies: A Qualitative Inquiry. AI and Tech in Behavioral and Social Sciences, 1(1), 4–11. https://doi.org/10.61838/kman.aitech.1.1.2

Khan, A. (2023, July 25). I tried an AI personal trainer so you don’t have to. The Verge. https://www.theverge.com

Novatchkov, H., & Baca, A. (2013). Artificial intelligence in sports on the example of weight training. Journal of Sports Science and Medicine, 12(1), 27–37.

Plumb, R. (2022, April 5). Peloton Guide review: Smart home gym gets smarter. TechCrunch. https://www.techcrunch.com

Ritchey, C. (2024, November 8). I Let ChatGPT Train Me for a Month. Here’s What Happened. Men’s Health. https://www.menshealth.com/fitness/a62629358/ai-personal-training-fitness/

WHOOP. (2023, September 26). WHOOP Unveils the New WHOOP Coach Powered by OpenAI. https://www.whoop.com/us/en/press-center/whoop-unveils-the-new-whoop-coach-powered-by-openai/

Pull Quotes:

“Artificial intelligence (AI) has rapidly penetrated deep into the fitness industry.”

“A primary duty of a strength coach is to develop a personalized training program tailored to an individual’s goal.”

“Building strength requires motivation to adhere to a rigorous training regimen.”