Facebook's accessibility team is working on a technology that will be able to describe photos for the visually impaired, which it hopes to release by the end of 2015.
Up to now, there has been technology available that can "read" the content on some social media sites, but the artificial intelligence tool the Facebook team is working on is completely new.
The team is lead by Matt King, Facebook’s first blind engineer.
A Facebook spokesperson says the tool will use similar deep learning techniques that face recognition technology Facebook already uses to identify individuals in photographs. He says the technology will describe photos, then feed those descriptions into a computer’s text-to-speech tool.
There are at least 50,000 blind people, and many more who are visually impaired who currently use Facebook, using text-to-speech technology such as Apple Voiceover.
Internet experts say the internet's heavy demand and reliance on vision has lead to many concerns about its accessibility for people with disabilities. They cite the ever increasing popularity of Instagram, which leaves vision-impaired users on the outside. They say although the visually impaired can take photos and post them to the social network, they are not able to 'view' photos and have to rely on the captions written by the posters.
Some tools that help take the internet from a visual to an auditory interaction already exists with text-to-speech programs even being able to interpret emojis, which for example use descriptions like “grinning face with squinting eyes.” But an impaired person's experience of an emoji can only be as good as the captions themselves.
After four years in development, Facebook hopes to release the tool by the end of 2015 to at least one platform, either web or iOS, and allow people to opt in to experience it.
Yesterday, Facebook boss Mark Zuckerberg addressed a Town Hall Q&A at the Indian Institute of Technology in New Delhi, "A.I. is a really exciting area of development. We have this goal in five to 10 years, we want to build computer systems, which can be better at main human senses than people are — can see better, recognize people or things in the world, can kind of track as we travel through the world, can hear better, can translate language better, can understand the language, right? All these things are basic human senses.”