Artificial intelligence (AI) is increasingly being explored for its applications in personal guidance, including the development of exercise programs and general goal setting. While experts identify potential benefits such as reducing barriers to self-reflection and offering accessible alternatives to human coaches, concerns persist regarding the accuracy, comprehensiveness, and potential biases of AI-generated advice. Recommendations for optimizing AI use emphasize providing detailed information, critically analyzing responses, and verifying sources.
Emergence of AI in Personal Guidance
The application of artificial intelligence in developing personalized programs and assisting with goal setting represents an evolving area of interest. A report from OpenAI indicated its AI chatbot is frequently utilized as an advisor. Dr. Keith Diaz, an associate professor at Columbia University Medical Center, described initial research into AI for fitness as "early phase."
Potential Applications and Benefits
AI systems are identified as potentially beneficial across various aspects of personal development and exercise:
- Goal Initiation and Structuring: AI can reduce barriers to self-reflection, empower individuals, and serve as a scaffold for expressing and understanding ideas. Ziang Xiao, an assistant professor at Johns Hopkins University, notes that AI can efficiently synthesize information to aid in organizing thoughts for goal initiation.
- Exercise Program Development:
- Starting Point: AI can serve as an initial resource for individuals beginning an exercise routine.
- Alternative Support: For those without access to in-person coaches, AI can assist in program development.
- Routine Modification: Experienced exercisers can use AI to vary or refresh existing routines, potentially generating novel workout suggestions.
- Motivation and Reinforcement: With appropriate prompts, AI platforms can highlight the importance of physical activity and suggest diverse engagement methods.
Risks and Limitations
Experts identify several limitations and risks associated with relying on AI for personal and exercise guidance:
- Bias and Generality: Large language models, trained on human-generated data, can reproduce assumptions and biases, potentially leading to generic or culturally specific goals that may not align with an individual's true desires. Chatbots may also incorporate irrelevant information despite previous interactions.
- Accuracy and Comprehensiveness: A 2024 study in JMIR Medical Education found AI-generated exercise recommendations to be approximately 90% accurate regarding facts but only about 40% comprehensive.
- Persuasion and Misdirection: AI chatbots can be persuasive, potentially guiding users toward mismatched or inappropriately affirmed goals, and may prioritize agreement over accuracy. Users with less technical expertise may be more susceptible to incorrect or harmful responses.
- Safety Concerns for Specific Populations: Dr. Laura A. Richardson, a clinical associate professor at the University of Michigan, notes that AI may not be suitable for individuals recovering from cardiac issues, other physical injuries, or those with low fitness levels. A November Circulation advisory from the American Heart Association (AHA) also noted limits in AI safety evaluations for healthcare tools.
- Inability to Replace Human Oversight: AI cannot replace medical professionals or provide real-time monitoring, physical execution feedback, or human motivation. It cannot detect signs of fatigue or the need for immediate modifications to workout intensity.
- Errors and Lack of Verification: AI programs are prone to making mistakes, and generated references may be non-existent or inaccurate.
Richardson suggests AI is most helpful for individuals in the middle of the fitness spectrum, while beginners with health conditions and elite athletes typically require higher levels of human oversight.
Guidelines for Optimizing AI Advice
To obtain the most relevant and accurate advice from AI, experts recommend:
- Detailed Input: Providing specific information, such as current exercise habits, maximum lift weights, target muscles, available equipment, other physical activities, and goals (e.g., for the next 3-6 months), generally yields more tailored advice.
- Role Assignment: Instructing the AI to act as a specific role, such as a "personal trainer" or "weight loss coach," can influence the quality of the response.
- Critical Analysis and Feedback: Critically analyze chatbot responses and provide informative feedback to improve personalization and realism.
- Goal Brainstorming: Prompt AI to brainstorm actionable goals, identify potential obstacles, and develop backup plans.
- Progress Tracking: Utilize AI as a collaborator for tracking progress.
- Self-Reflection: Be wary of tools that bypass self-reflection, and consider reviewing past reasons for unachieved goals to understand underlying priorities.
- Reference Verification: Request references from AI and independently verify their existence and content.
Overall Perspective
Dr. Diaz suggests that AI's potential in promoting physical activity likely outweighs potential harms, provided clear guidance on appropriate usage is developed. Richardson underscores that exercise technology, including AI, serves as a tool rather than the ultimate goal, advising a focus on the physical and mental benefits of movement. Ultimately, AI chatbots can function as reflective partners, but responsibility for actions and outcomes remains with the user.
For general health, the American Heart Association recommends that adults engage in at least 150 minutes per week of moderate-intensity aerobic exercise, 75 minutes per week of vigorous aerobic exercise, or a combination. Additionally, muscle-strengthening activities of moderate to high intensity are advised at least two days per week for those able.