Artificial Intelligence
AI Restores Vision in Patients With Irreversible Blindness
In a new study, AI helps restore vision to patients with untreatable blindness
Posted October 21, 2025 Reviewed by Margaret Foley
A groundbreaking study published in The New England Journal of Medicine (NEJM) on Monday demonstrates how specialized devices using artificial intelligence can help restore central vision to treat what was previously a common form of untreatable blindness. This revolutionary technology solution paves the way toward reestablishing sight capabilities for those with vision loss due to geographic atrophy from age-related macular degeneration (AMD).
To achieve this remarkable medical milestone, a global team of researchers led by José-Alain Sahel, M.D., at the University of Pittsburgh School of Medicine, Daniel Palanker, Ph.D., professor of ophthalmology at the Byers Eye Institute at Stanford, and Frank Holz, M.D., at the University of Bonn, created cutting-edge hardware combined with an AI software solution to restore vision to those who have lost central vision and have geographic atrophy due to age-related macular degeneration.
Globally, age-related macular degeneration is among the most common causes of blindness and vision impairment, along with cataracts, glaucoma, diabetic retinopathy, and uncorrected refractive errors, according to the World Health Organization. By 2050, an estimated over 1.76 billion people of all ages globally will be living with vision loss, consisting of roughly 980 million females and 781 million males, according to a study published in The Lancet Global Health by Bourne R., et al.
Age-related macular degeneration impacts the center of the retina, called the macula, and is characterized by impaired central vision and the loss of detailed vision. Central vision is the ability to see straight in front of you and is critical in performing daily life tasks such as driving, watching a movie, viewing a smartphone, checking social media posts, viewing videos, and reading. Unlike peripheral vision or indirect vision, central vision is typically sharper because your eyes are usually focused on what they are viewing.
There are two forms of age-related macular degeneration: dry and wet. Dry age-related macular degeneration, also known as geographic atrophy, is the most common form of AMD that comprises 80 percent of cases, according to the American Academy of Ophthalmology. Geographic atrophy happens when areas of the retina, especially the macula, start to thin out and eventually stop working altogether, which may lead to the gradual loss of central vision, blurry vision, distorted sight, dulling of colors, challenges seeing in low light conditions, and issues with recognizing faces. The American Academy of Ophthalmology estimates there are 8 million people globally with geographic atrophy.
This breakthrough study included human participants from 17 sites in multiple countries, including the United Kingdom, France, Italy, Germany, and the Netherlands. From an original pool of 38 study participants, 32 were evaluated at 12 months and had clinically meaningful improvements in the sharpness and clarity of vision.
At the core of the solution is an innovative wireless retinal implant solution called PRIMA, short for photovoltaic retina microarray, which was developed by Palanker at Stanford University. The three main parts of the PRIMA solution include a tiny wireless implant, augmented reality-like video glasses with a built-in camera, and a small processor that fits in a pocket that processes visual data using an AI algorithm.
To install the retinal implant in the eye, the gel-like, clear substance that fills the eye, called the vitreous humor, is removed. This procedure is called a vitrectomy. Next, an ultra-thin, 30-micron chip containing 378 electrodes in a minuscule space of just 2x2 millimeters, is placed under the macula in the area impacted by atrophy. The macula is a circular area located in the middle of the retina at the back of the eyeball.
The way the solution works is that the glasses take in the view in real time and project it to the implanted chip in the eye by pulsing near-infrared (NIR) light with information about the image. The implant contains photovoltaic cells that convert light into electricity. In other words, the retinal implant transforms the pulses of light received from the glasses into pulses of electric current in each pixel to excite the retinal neurons.
A pocket-sized computer uses AI to process the data in order to transform it into electrical signals that travel through the retina and nerve cells of the eye in order to reach the brain, which translates it into vision. Artificial intelligence machine learning is often used to detect patterns in noisy data. In this study, the pattern-recognition and predictive capabilities of AI are used to convert pulses of near-infrared light into electrical current.
In effect, machine learning is helping to bridge the gap between the raw data of the visual scene captured by the integrated camera in the special glasses into a form that can assist the brain in interpreting what image the translated data represents.
“In this study involving 38 participants with geographic atrophy due to AMD, the PRIMA system restored central vision and led to a significant improvement in visual acuity from baseline to month 12,” the scientists reported.
A majority of the study participants (84 percent) were able to read numbers, letters, and words from what used to be a blind eye due to the PRIMA system, according to a University College London report.
Harnessing the powerful capabilities of AI machine learning with innovative hardware devices may be an effective way to restore vision for one of the leading causes of irreversible blindness. This pioneering research is paving the way for those suffering from irreversible blindness to see once again in the future.
Copyright © 2025 Cami Rosso. All rights reserved.
