Eighteen months ago, I suffered a stroke that placed me among the 24% of Americans living with chronic pain. The stroke left me with persistent, low-level pain affecting most of the right side of my body but fortunately spared my cognitive capacities. As an economist deeply interested in artificial intelligence, I quickly turned to AI to find ways of reducing my pain and have achieved meaningful success with interventions my medical providers had not suggested, particularly green light therapy and compression clothing. Finding these successful interventions was not straightforward and required significant use of AI.

Many individuals suffering from chronic pain lack either the technical skills or cognitive capacity required to effectively navigate LLMs independently. Meanwhile, physicians often do not possess comprehensive knowledge of the vast array of available non-pharmaceutical interventions. There is a significant unmet need for a simple, accessible, and entirely free AI-powered tool specifically designed for chronic pain management.

I propose the creation of either a charity-based AI wrapper explicitly dedicated to chronic pain management, or suggest that an existing AI firm create such a wrapper. This wrapper would ideally remain completely free to users as charging even a minimal fee would discourage many from trying it. It should also be established as a separate legal entity to provide liability protection, and anyone pursuing this project will need to consult with a qualified legal professional.

Why Chronic Pain Matters

The impact of chronic pain extends beyond physical discomfort.  Between 20% and 50% of chronic pain patients meet criteria for depression.

Economically, chronic pain is extraordinarily costly. In the US alone, the cost of lost worker productivity associated with chronic pain was estimated to range from $560 to $635 billion per year. 

Despite the magnitude of this burden, chronic pain research is underfunded. Analyses of National Institutes of Health grants indicate that roughly 1% of the NIH budget was directed to pain research in 2003.  For a condition that disables millions and costs hundreds of billions of dollars, this funding allocation implies significant neglect.

Addressing pain effectively would recover healthy life years, reduce depression and anxiety, and alleviate a large economic drag. Given its scale and neglect, chronic pain deserves recognition as a central global health priority.

Why Current AI Alone Isn’t Good Enough
Based on my experience, current AI tools, such as ChatGPT, remain insufficient on their own for effectively managing chronic pain:

• Poor at spontaneously suggesting treatments: AI typically does not propose novel interventions unless explicitly prompted, making it challenging for users unfamiliar with the medical landscape to independently discover helpful solutions.
• Limited practical guidance: LLMs rarely suggest affordable and easily obtainable products, nor do they clearly indicate costs or purchasing options.
• Challenges with accurate medical input: Optimal recommendations require detailed medical inputs (age, sex, previous health conditions, exact injury location). Current AI tools are not designed to guide users through entering this information.
• AI hallucinations are minefields for those not able to verify information.

A dedicated AI wrapper specifically designed for chronic pain management could resolve these limitations, providing enhanced usability, safety, accuracy, and practical recommendations well beyond what standalone AI currently offers.

Vision for a Charity-Based AI Pain Management Wrapper
An ideal AI wrapper would function as follows:

• Doctor-Led Medical Information Input: Ideally, a medical professional would directly input comprehensive patient data including exact medical diagnosis (e.g., stroke location in brain), age, sex, medical history, and current medications before referring the patient to the AI program or app. However, the program should not rely solely on this input method and should provide options for users or caregivers to input this information themselves. The AI would account for an increased possibility of error when information is not entered by a medical professional and exercise greater caution if an intervention could be harmful under slightly different medical circumstances.
• Adaptive Sophistication Detection: The AI would quickly evaluate each patient's familiarity with medical terminology and statistical concepts, automatically adjusting its language and communication style to optimize ease of understanding. Based on my personal experience, I have been able to convince the AI that I understand statistics at a high level, but it keeps using medical terminology I do not understand.
• Conversational Voice-Based Accessibility: The AI should interact through multiple formats, including voice, a phone app, or a computer-based interface.
• Personalized Treatment Recommendations: The wrapper would match each user's profile and condition to interventions. For high-risk interventions the AI would suggest the user talk to a doctor rather than provide a way to implement the recommendation. For low-risk ones the AI would provide direct links to (ideally) affordable products.
• Try, Evaluate, and Iterate: The program would present multiple tailored recommendations at a manageable pace, encouraging users to try one or two interventions at a time, evaluate their effectiveness, and continue iteratively exploring additional treatments until optimal relief is achieved.
• Daily Conversational Support and Monitoring: Regular, brief conversations would reinforce adherence, promptly resolve issues, and allow the AI to adapt its recommendations. Periodic validated pain assessments would quantify treatment effectiveness and support ongoing research.
• Entirely Free and Easy Adoption: Critically, this wrapper should be entirely free to ensure widespread adoption. Charging fees, even minimal ones, discourages usage and complicates accessibility. Even a nominal fee would require users to establish trust and manage payment processes.

• Continuous Evolution: As AI technology improves, the wrapper would continuously adapt and ultimately become obsolete once standalone AI sufficiently addresses these needs. Until that point, it would provide significant incremental value.

   

Existing AI-Driven Pain Management Efforts
Several commercial and nonprofit initiatives already integrate AI into pain management:

• Commercial companies (e.g., Kaia Health, Hinge Health, SWORD Health) offer AI-driven posture analysis and targeted exercises primarily for musculoskeletal conditions, typically through employer or insurer plans.
• Subscription-based apps (Curable, Pathways) deliver general pain education and mindfulness, usually charging $10–15 per month, creating affordability barriers.

To the best of my knowledge, no one offers fully personalized medical integration, accessible voice-driven interaction, detailed product guidance, extensive non-pharmaceutical recommendations, and completely free usage.

Possible Pain Treatments

The set of pain treatments is sufficiently vast so that you cannot expect any health care providers to be familiar with all of them.  The set includes: Acceptance and Commitment Therapy, Acupressure mat, Acupuncture, Alexander Technique, Animal-assisted therapy, Aromatherapy (e.g., using essential oils like lavender), Art therapy, Balneotherapy, Biofeedback, Breathing exercises, Chiropractic care, Cognitive behavioral therapy, Cold therapy, Counterirritants, Cranial electrotherapy stimulation, Craniosacral therapy, Cupping therapy, Dance/Movement therapy, Deep brain stimulation, Dietary modification, Distraction techniques, Dry needling, Emotional Awareness and Expression Therapy, Ergonomic adjustments, Extracorporeal shockwave therapy (Shockwave Therapy), Feldenkrais method, Floatation tank, Graded exposure therapy, Graded motor imagery, Green light therapy, Grounding, Guided imagery, Heart rate variability training, Heat therapy, Humor/Laughter therapy, Hydrotherapy, Hyperbaric oxygen therapy, Hypnosis (or hypnotherapy), Interferential current, Journal writing, Kinesiology taping, Laser therapy (low-level laser therapy), Magnet therapy (static or electromagnetic), Massage, Meditation, Mindfulness-Based Stress Reduction, Mirror therapy, Multidisciplinary rehabilitation, Music therapy, Neurofeedback, Noninvasive vagus nerve stimulation, Occupational therapy, Osteopathic cranial manipulation, Osteopathic manipulative treatment, Over-the-counter supplements, Pain reprocessing therapy, Patient education and self-management programs, Peripheral nerve stimulation, Physical therapy, Positioning, Progressive muscle relaxation, Radiofrequency ablation (radiofrequency neurotomy), Red light therapy, Reflexology, Reiki, Sauna, Scrambler therapy, Sleep optimization, Social support groups, Spinal cord stimulation, Spirituality and religion (e.g., prayer or faith-based practices), Sunlight exposure, Tai Chi or Qigong, Therapeutic spinal traction/inversion therapy, Therapeutic ultrasound, Transcranial direct current stimulation, Transcranial magnetic stimulation, Transcutaneous electrical nerve stimulation, Vibration therapy, Virtual reality or extended reality, Walking, Weight management, Yoga or stretching.

Call to Action
I openly share this proposal and explicitly disclaim any personal intention of pursuing it.  Chronic pain remains an underserved, high-impact issue precisely aligned with Effective Altruism’s priorities. Creating this charity-based AI wrapper represents a concrete opportunity to significantly reduce suffering worldwide.