Artificial intelligence in the eye care sector

Share article
Author
DSc* MSc* FCOptom FAAO Wolfgang Cagnolati1
1*Pennsylvania College of Optometry at Salus University, Philadelphia, PA, USA, Berliner Hochschule für Technik (BHT), University of Applied Sciences, Berlin, Germany, Optometrie Cagnolati GmbH, Duisburg, Germany

In 1950, Alan Turing was the first to describe his thoughts on the use of computers for artificial intelligence (AI) in his book “Computers and Intelligence.”1 The goal of AI is to mimic human cognitive abilities. In their interesting publication “History of artificial intelligence in medicine,” Kaul et al. describe both the development of AI and its introduction and advancement in medicine over the past decades.1

The number of scientific and clinical publications on AI in ophthalmology and optometry already demonstrates its enormous potential across the entire eye care sector. This is particularly true in the field of ophthalmic imaging and its evaluation.2 In addition to traditional AI-based examinations by an optometrist or ophthalmologist, it is now also possible to use smartphone imaging in conjunction with AI technology, especially for diagnostics on the anterior segment of the eye. 3 Nevertheless, the question of integrating AI into the field of medicine has not yet been sufficiently clarified.4 This also includes selective screening for individual eye diseases based on AI by non-specialists. If we look at the uncontrolled growth of service offerings in this context, which are appearing on the “market” almost daily, it is clear that there is an enormous need for regulation.

AI is also already being used in training and computer-assisted eye examinations. Stuermer et al. describe in a recent publication how a virtual assistant can be used with what level of accuracy in primary eye care.5 This AI-based virtual assistant was trained using tabular clinical data with the aim of being used both in everyday clinical practice and in training programs for optometrists. The following three models were trained using machine learning algorithms:

  • Model 1 for predicting the most likely case classification
  • Model 2 for predicting pathology in different eye segments
  • Model 3 for predicting any type of binocular vision disorder

The results were excellent.

The possibilities of AI in the whole eye care sector are already enormous today and will increase exponentially in the near future. Patient safety in the context of possible false positive or false negative findings, but also the protection of personal data, is one of the central challenges for the use of AI in healthcare. Responding adequately to this is an important task for the associations of both eye care professions, but also for the regulatory authorities.

References

[1] Kaul, V., Enslin, S., Gross, S. A. (2020). History of artificial intelligence in medicine. Gastrointest. Endosc., 92, 807-812. 

[2] Waisberg, E., Ong, J., Kamran, S. A., Masalkhi, M., Paladugu, P., Zaman, N., Lee, A. G., Tavakkoli, A. (2025). Generative artificial intelligence in ophthalmology. Surv. Ophthalmol., 70, 1-11.

[3] Oshika, T. (2024). Artificial Intelligence Applications in Ophthalmology. JMA J., 8, 66-75. 

[4] Ong, J. C. .L, Chang, S. Y., William, W., Butte, A. J., Shah, N. H., Chew, L. S. T., Liu, N., Doshi-Velez, F., Lu, W., Savulescu, J., Ting, D. S. W. (2024). Ethical and regulatory challenges of large language models in medicine. Lancet Digit. Health, 6, e428-e432.

[5] Stuermer, L., Braga, S., Martin, R., Wolffsohn, J. S. (2025). Artificial intelligence virtual assistants in primary eye care practice. Ophthalmic Physiol. Opt., 45, 437-449.