Artificial intelligence transforming glaucoma detection and care
AI now surpasses expert clinicians in glaucoma screening while expanding telemedicine access and reshaping how ophthalmologists communicate with patients about eye health.
How artificial intelligence is changing glaucoma detection
Glaucoma remains one of the leading causes of irreversible blindness worldwide, yet many people don’t realize they have it until significant vision loss occurs. The disease progresses silently in its early stages, making early detection crucial. Now, artificial intelligence is emerging as a powerful tool that’s fundamentally changing how ophthalmologists identify and manage this sight-threatening condition.
At the 44th Congress of the ESCRS in London, ophthalmologist Ana Miguel, MD, PhD, FEBO, FEBO-G, GCSRT, discussed groundbreaking research demonstrating that AI systems now outperform even expert glaucoma specialists in screening. Her team published findings in The Lancet comparing AI algorithms directly against a group of European glaucoma experts, and the results were striking: the AI system demonstrated superior detection rates.
The study results: AI beating the experts
The research revealed something that initially surprised many clinicians—machines trained on large datasets could identify glaucomatous changes more consistently than human experts. This doesn’t mean ophthalmologists should feel threatened. Instead, Miguel emphasized that clinicians need to embrace this technology as a collaborative tool rather than view it as competition.
The practical benefit extends beyond individual patient care. Because AI can handle much of the screening workload, it frees up valuable expert time. This means more patients can receive comprehensive evaluations overall. In Miguel’s words, doctors need to “use it cleverly” while still relying on their own clinical judgment, what she calls “natural intelligence.”
Expanding access through telemedicine screening
Miguel has integrated AI extensively into telemedicine programs, particularly in regions where healthcare access is limited. In parts of France where she practices, the telemedicine model works like this: orthoptists and trained support staff see patients, performing retinal photography, refraction measurements, and intraocular pressure testing using air-puff tonometry.
An AI algorithm then reviews these images and measurements to determine whether a patient needs a full in-person consultation with an ophthalmologist. Importantly, Miguel personally reviews every single retinal photo herself—maintaining that critical human verification step. This hybrid approach allows her to screen more patients without compromising quality, since the algorithm helps prioritize cases that truly need expert evaluation.
This model is particularly valuable in underserved areas where ophthalmologists are scarce. Patients can receive initial screening and risk assessment locally, with only those requiring specialist care traveling for consultation. It’s a practical solution to healthcare access problems that affect millions globally.
Real barriers still exist
Despite AI’s promise, significant obstacles slow wider adoption. Fear and skepticism remain common, particularly among older clinicians and those trained in more traditional practice approaches. This resistance isn’t entirely unreasonable—poorly trained or inadequately validated AI systems can produce dangerously wrong conclusions.
Rigorous validation and regulatory oversight are essential before any AI tool enters clinical practice. Miguel stressed that clinical validation is non-negotiable. Each system must be tested thoroughly, and importantly, human verification serves as a critical second check. An algorithm might spot subtle changes humans miss, but a knowledgeable ophthalmologist must confirm the findings before treatment decisions are made.
The technology itself isn’t the only hurdle. Integration into existing clinical workflows, training requirements for staff, and questions about liability and responsibility all complicate implementation. Organizations considering AI adoption must address these practical concerns alongside the scientific validation.
Patients are already using AI—whether doctors like it or not
A significant shift is already happening in the patient-doctor dynamic. Younger patients especially are turning to AI tools like ChatGPT to research their symptoms and eye health concerns. This trend creates both opportunities and challenges.
ChatGPT and similar tools have evolved their responses over time. Rather than offering definitive diagnoses, these systems now emphasize that patients should see a physician for proper evaluation. Miguel views this shift as having both positive and negative aspects. On one hand, patients become more engaged in understanding their own health. On the other hand, the same trend can undermine patient confidence in their ophthalmologist.
This highlights why the doctor-patient relationship remains irreplaceable. When patients trust their ophthalmologist and feel heard, they’re more likely to be honest about symptoms and more adherent to treatment plans—especially important for glaucoma management, where consistent use of eye drops is crucial for controlling intraocular pressure.
The future requires adaptation
Miguel’s clear message to her colleagues is straightforward: “AI is absolutely here to stay, and you have to adapt to it.” Resistance is futile; the technology will continue advancing and expanding into clinical practice whether reluctant practitioners embrace it or not.
The question isn’t whether to use AI, but how to use it responsibly and effectively. Ophthalmologists who understand AI’s capabilities and limitations—and who maintain strong clinical oversight—will be best positioned to offer patients the most comprehensive care. The goal isn’t to replace expert judgment but to enhance it, catching more cases of glaucoma earlier when treatment can prevent vision loss.
For patients with glaucoma or at risk for the disease, this transformation means more screening opportunities, faster detection, and better coordination between specialists and primary care providers. For ophthalmologists, it means learning new tools and adapting workflows, but ultimately practicing more effective medicine.
When to schedule a glaucoma screening
If you haven’t had a comprehensive eye exam in the past two years, schedule one now—especially if you’re over 40, have a family history of glaucoma, have high intraocular pressure, or are of African descent (which carries higher glaucoma risk). Early detection through screening, whether assisted by AI algorithms or traditional methods, remains your best defense against vision loss from glaucoma.
