For many people, accessibility settings are just a place to adjust your phone's text size. But for the tens of thousands of people with impaired vision who use voiceover software and text-to-speech tools, they're the single most important portal to the 21st century.
Now, Facebook is using artificial intelligence and image recognition software to help people with visual impairments get access not just to the text on their screens but the images as well. Instead of simply reading image titles and caption text, Facebook's new artificial intelligence will scan an image, recognize its elements and describe, in a few keywords, the contents of that photo.
Facebook has an entire team dedicated to this project. Before, the best options for people with visual impairments to use Facebook were voice-to-text programs that simply read the text and entry fields top to bottom; about 50,000 people use Facebook using a program called Apple Voiceover. This team is looking for new ways, tools and controls to bring Facebook to those who aren't often considered in the design and creation of contemporary phones and apps.
"I mean, it's the gateway to employment, it's the gateway to opportunity of all different kinds — participating in your government and everything," Matt King, Facebook's first blind engineer, told TechCrunch. "So when we can flatten that on-ramp, I see that as the ultimate goal in accessibility and I think that Facebook is really uniquely positioned to help do that. ... It's a way of giving dignity to every person with a disability in the world by helping them get connected to everybody else."