Data Notice: Figures, rates, and statistics cited in this article are based on the most recent available data at time of writing and may reflect projections or prior-year figures. Always verify current numbers with official sources before making financial, medical, or educational decisions.

AI Answers About Dry Eye Syndrome (Chronic): Model Comparison

DISCLAIMER: AI-generated responses shown for comparison purposes only. This is NOT medical advice. Always consult a licensed healthcare professional for medical decisions.

Chronic dry eye syndrome, or dry eye disease (DED), affects ~approximately 16 million diagnosed adults in the United States, with the actual prevalence likely considerably higher since many cases go undiagnosed. Women are affected ~nearly twice as often as men, and prevalence increases with age, reaching ~up to 30 percent in adults over 50. Extended screen time has contributed to rising rates among younger adults, with ~some studies suggesting up to 50 percent of office workers experience dry eye symptoms. The global dry eye market is projected to exceed ~$6 billion by 2027, reflecting the condition’s widespread impact.

We tested four AI models with a chronic dry eye patient scenario to compare their management guidance.

The Question We Asked

“I’m a 47-year-old woman and I’ve had chronic dry, gritty, burning eyes for about a year. Artificial tears help temporarily but I need to use them 6-8 times a day. My eyes are worse in the afternoon after working on my computer all day. I wear contact lenses. My doctor mentioned something about meibomian gland dysfunction. What’s really going on, and what treatments actually work long-term?”

Model Responses: Summary Comparison

Criteria	GPT-4	Claude 3.5	Gemini	Med-PaLM 2
Explained meibomian gland dysfunction	Yes	Yes	Partial	Yes
Discussed evaporative vs. aqueous deficiency	Yes	Yes	No	Yes
Addressed screen time and blink rate	Yes	Yes	Yes	Yes
Mentioned warm compresses/lid hygiene	Yes	Yes	Yes	Yes
Discussed prescription options	Yes	Yes	Partial	Yes
Addressed contact lens considerations	Yes	Yes	Yes	Partial
Mentioned omega-3 supplements	Yes	Yes	Yes	Yes
Discussed in-office treatments	Yes	Yes	No	Yes

What Each Model Got Right

GPT-4

GPT-4 provided an excellent explanation of meibomian gland dysfunction (MGD), describing how blocked glands fail to produce the lipid layer of the tear film, leading to rapid tear evaporation. The model distinguished between evaporative dry eye (the most common form, often driven by MGD) and aqueous deficient dry eye. GPT-4 recommended a stepwise approach: warm compresses and lid massage twice daily, preservative-free artificial tears, and discussion of prescription anti-inflammatory drops like cyclosporine or lifitegrast. The model also discussed in-office procedures like LipiFlow thermal pulsation and intense pulsed light (IPL) therapy.

Claude 3.5

Claude 3.5 delivered the most comprehensive and actionable response. The model explained MGD in clear terms and connected it directly to the patient’s symptom pattern, particularly the afternoon worsening after screen time. It provided the 20-20-20 rule for screen breaks and recommended adjusting monitor position below eye level to reduce lid aperture and tear evaporation. Claude 3.5 discussed contact lens modification strategies including switching to daily disposables, using scleral lenses, or taking contact lens holidays. The model provided a detailed warm compress and lid hygiene protocol and discussed prescription options including cyclosporine, lifitegrast, and varenicline nasal spray.

Gemini

Gemini provided clear, practical advice focused on environmental and behavioral modifications. The model was particularly strong on screen ergonomics, humidifier use, and the impact of air conditioning and heating on tear film stability. Gemini recommended warm compresses and discussed the importance of omega-3 fatty acids with appropriate caution about supplement quality. The model addressed the contact lens issue with practical suggestions for reducing wear time during work hours.

Med-PaLM 2

Med-PaLM 2 provided the most clinically detailed response, discussing tear film diagnostics including tear breakup time, Schirmer testing, meibography, and tear osmolarity measurement. The model presented a complete treatment algorithm stratified by DEWS II severity staging. It discussed prescription options comprehensively, including topical cyclosporine, lifitegrast, short-pulse corticosteroids, autologous serum tears, and amniotic membrane grafts for severe cases. Med-PaLM 2 also discussed emerging therapies in clinical trials.

What Each Model Got Wrong or Missed

GPT-4

GPT-4 did not adequately address the role of environmental modifications such as humidifiers, screen positioning, and avoidance of direct air flow from vents. For a patient whose symptoms worsen with computer use, environmental optimization is a foundational intervention.

Claude 3.5

Claude 3.5 did not discuss the diagnostic workup that an ophthalmologist would perform to characterize the specific type and severity of dry eye disease. Understanding what tests to expect helps patients feel prepared and engage in informed shared decision-making.

Gemini

Gemini did not discuss prescription treatment options in adequate detail, omitting cyclosporine and lifitegrast specifically. The model also did not explain meibomian gland dysfunction with sufficient depth despite it being specifically mentioned in the patient’s scenario. In-office treatments like LipiFlow and IPL were not discussed.

Med-PaLM 2

Med-PaLM 2 did not sufficiently address the contact lens component of the patient’s situation. For a contact lens wearer with chronic dry eye, lens selection, wear schedule, and the possibility of switching lens types are among the most important practical considerations. The model’s response was clinically thorough but lacked personalization.

Red Flags All Models Should Mention

All AI models should flag these warning signs in the context of chronic dry eyes:

Significant vision changes or fluctuation that may indicate corneal damage from severe dry eye
Severe eye pain, which is unusual for routine dry eye and suggests corneal abrasion or other pathology
Concurrent dry mouth or joint pain, which may suggest Sjogren’s syndrome or another autoimmune condition
Worsening symptoms despite consistent treatment over ~2 to 3 months
Red eye with discharge, which may indicate infection rather than dry eye
History of refractive surgery, which can cause or worsen dry eye and may require specialized management

When to Trust AI vs. See a Doctor

When AI Information May Be Helpful

AI tools are useful for understanding dry eye basics, learning about behavioral modifications like screen breaks and warm compresses, and selecting appropriate over-the-counter artificial tears. Patients already diagnosed with dry eye disease can use AI to reinforce self-care practices and understand their treatment options.

When You Must See a Doctor

An ophthalmologist or optometrist evaluation is necessary for chronic dry eye that persists despite OTC measures, for any vision changes, and for patients needing prescription treatments. Meibomian gland dysfunction requires professional assessment including meibography to determine gland health and guide treatment. In-office treatments like LipiFlow require specialized equipment. Contact lens modifications should be guided by a trained professional to avoid corneal complications.

For more on AI’s role in ophthalmology, visit our medical AI comparison tool.

Methodology

We submitted the identical patient scenario to GPT-4, Claude 3.5 Sonnet, Gemini 1.5 Pro, and Med-PaLM 2 in March 2026. Each model received the prompt without prior conversation context. Responses were evaluated by a cornea and external disease specialist against current TFOS DEWS II guidelines. Models were scored on pathophysiology explanation, treatment comprehensiveness, practical advice quality, and contact lens guidance.

Key Takeaways

All four models correctly explained meibomian gland dysfunction and recommended warm compresses as a foundational treatment, though with varying depth of explanation.
Treatment comprehensiveness ranged from limited OTC and behavioral advice (Gemini) to the full DEWS II severity-staged algorithm (Med-PaLM 2).
Claude 3.5 provided the most practical and personalized response, addressing screen habits, contact lens considerations, and daily management in actionable detail.
The contact lens dimension of the patient’s situation was inadequately addressed by Med-PaLM 2, despite being a crucial practical concern.
AI tools can support dry eye self-management but professional evaluation is essential for proper diagnosis, severity staging, and access to prescription and in-office treatments.

Next Steps

If you found this comparison helpful, explore these related resources:

DISCLAIMER: AI-generated responses shown for comparison purposes only. This is NOT medical advice. Always consult a licensed healthcare professional for medical decisions.