CSR: How AI could give sight to the blind

Better medical care, nutrition, and other measures can help prevent blindness or ease the suffering of those afflicted by it. Alongside these solutions, AI has the potential to make life easier for the visually impaired by providing navigational assistance via smartphone.

Applications

Artificial Intelligence’s computer vision capabilities can identify objects and read text, converting handwriting or printed text to digital text that can then be spoken aloud. Existing mobile applications include Microsoft’s Seeing AI. Reviews of the application indicate that it has very large daily applicability.

Similar solutions include the OrCam MyEye camera, which is mounted on standard glasses and converts what is seen into spoken output. The device is easily portable and operates without the need for a smartphone.

While advances in computer vision have improved scene description technology, this remains a developing capability. As it improves, it will provide visually impaired people with a richer and deeper understanding of their environment, including recognizing other people and describing colours, for example.

Potential barriers

Limited access to technology in emerging economies is a potential barrier that will need to be overcome. One of the most significant barriers to AI in this area is “last mile” implementation in the form of access to technology.

Many people do not own the smartphone needed for existing applications; smartphone penetration globally is below 40%. Building partnerships with NGOs and governments to provide basic technology access to individuals in poor communities could help address this limitation.

Talent is also a potential bottleneck. Developing AI applications to tailor them more closely to the needs of the visually impaired in various regions, including with more detailed description of the surrounding environment, may require the efforts of high-level AI researchers over a period of years.

Risks

Among the risks, data privacy looms large. Users of such a mobile application would capture images of the environment that often include other individuals or private data, for example pictures of credit cards, and may also rely on training data that could include private information.

Not being able to see is a major impediment to work. AI has tremendous potential for working age adults with visual impairment who are unemployed, to live and work with independence and dignity.

This article is part of a series on using AI for social good.

Subscribe