AI is Coming, Ready or Not

AI has the potential to significantly improve the quality of care in ophthalmology and provide better outcomes for patients. However, it is important to note that AI should be used as a tool to augment and support, rather than replace, human expertise and decision-making in ophthalmology.”

The preceding opinion was generated - word for word - by ChatGPT, the open AI online tool capable of producing detailed responses to any prompt based on what it finds online. I’m grateful and somewhat intrigued that a machine reminds us of the human element of what we do.

Telemedicine, robotics, machine learning, and artificial intelligence are profoundly changing countless elements of healthcare. Nowhere is this more true than in ophthalmology, where we are pioneering new ways to use information technology to increase safety and efficiency while improving outcomes.

We should embrace the wave of new digital tools steadily making their way into clinical practice and have the potential to greatly assist in patient care. As Howard Larkin points out in our cover article this month, we are entering the era of the digital OR. We are now seeing every step of cataract surgery begin to integrate data from patient records, preoperative imaging, AI databases, and manufacturers. How much will quality improve, and how much will it cost? This remains to be seen.

Another article in this issue features the work of Dr Flora Lum, who oversees the American Academy of Ophthalmology’s Intelligent Research in Sight (IRIS) Registry.

Her work shows the utility of large data sets for examining practice patterns, reporting quality, improving quality at individual practices, and advancing scientific discovery. Similar projects are underway using data from the European Registry of Quality Outcomes for Cataract and Refractive Surgery (EUREQUO).

Cataract surgery has already felt the impact of machine learning through more accurate biometry formulas for IOL power calculation. The recently launched ESCRS IOL Calculator compares many different IOL formula calculators (see https://iolcalculator.escrs.org).

In a related article, Dr Bruce Allan at Moorfields highlights the deficit in large data sets for anterior segment surgery. Recognizing the problem, the ESCRS recently announced a research call for a collection of open-access data sets. The goal is to create richly labelled data sets comprised of imaging and clinical outcomes data for groups of 1,000 to 100,000 patients.

We do need to temper our enthusiasm for AI by noting the potential risks and limitations—including concerns about cost, bias, bioethics, data protection, and patient privacy. As the chatbot reminds us, AI should be a tool to augment and support, not replace clinical decision-making.

Béatrice Cochener-Lamard MD, PhD is Professor and Head of the Ophthalmology Department at Brest University Hospital, France, and former President of the ESCRS.