Over the last few years, something big has happened to the way we communicate — though if you could see it happening, you may not have noticed it.
The proliferation of millions of connected cameras, along with the Internet’s general tendency toward loosening language have worked in tandem to nudge our day-to-day communications ever further away from the textual and ever more deeply into the visual. These days, a good portion of our talk takes the form of photos, videos, emoji, animated GIFs, and stickers; and in the process, social media has converted a significant portion of our social lives into consumable content — well, consumable for most of us.
This past week, Facebook made news with its unveiling of “automatic alternative text,” a technology that follows years of research, and stands to significantly enhance the service’s accessibility for the visually impaired, as well as the experience those users have.
Developed by Matt King, Facebook’s first blind engineer, automatic alternative text employs artificial intelligence to generate spoken interpretations of photos — so, a photo of two friends vacationing on the beach might be described through the voice of a screen reader program (like Apple’s VoiceOver or Microsoft’s Narrator) as showing two people, sand, water, and sky. It also reads aloud data like date, time, caption text, reactions, and comments. (Currently the feature is only available for iOS screen readers set to English, but Facebook plans to expand its avaiability across platforms and languages.)
The resulting experience is far richer and offers a deeper understanding (for both the visually impaired and Facebook) of what exactly people are posting.
Automatic alternative text arrives as part of a recent wave of advances in accessibility for the blind across tech and social media. Just last week, Twitter announced that iOS and Android users would now be able to add descriptions (i.e. old-fashioned non-automated “alt-text”) to any photo posted to the service. These text descriptions can extend up to 420 characters (no need to be pithy there), and are detected and read by screen reader programs. Twitter also opened up its API to accommodate alt-text descriptions through much relied-upon third-party Twitter clients for the visually impaired like EasyChirp, Chicken Nugget, and The Qube.
And these are just the beginning. As image and facial recognition capabilities grow more and more sophisticated, it’s easy to imagine ways that technology can go deeper. Microsoft recently launched a research initiative called Seeing AI, which aims to use artificial intelligence and a either a smartphone or a pair of Pivothead glasses to read everything from the sections of a menu to the moods of your co-workers).
Meanwhile, specialized GPS apps like the suite made by Sendero (motto: “It is better to travel hopefully than to arrive”) and indoor/outdoor navigational apps like BlindSquare are turning mobile phones into essential tools for the blind to experience life offline.
There’s still plenty to be done to make the Internet more accessible to the visually impaired. Automatic alt-text hasn’t yet been extended to Facebook’s narcissistic-little-sister-site Instagram, a platform that relies almost entirely on images, and remains frustratingly improvisational for the blind. And while the similarly image-driven site Tumblr offers the option to enter alt-text descriptions for photos, complaints have been common among users for years over how its design is often at odds with its accessibility. YouTube has made strides for its deaf users by further developing its sometimes reliable automatic captioning technology, but it still does not allow for users to upload audio descriptions for videos (which would require an extra audio track).
Accessibility isn’t just a matter of courtesy, it’s becoming a matter of legal contention. There’s been a surge of recent lawsuits alleging that websites and services that are inaccessible amount to violations of Title III of the Americans with Disabilities Act. And while it’s unlikely that the DOJ will offer hard and fast rules about whether websites counts as “public accommodations” (the statute was, after all, passed in 1990), the federal government already operates its digital communications to be compliant with Section 508, (which requires that “individuals with disabilities, who are members of the public seeking information or services from a Federal agency, have access to and use of information and data that is comparable to that provided to the public who are not individuals with disabilities”) by following a strict set of guidelines. It could be a forecast of what regulations might emerge in the near future.
Ideally, social media’s promise would extend equally to everyone — allowing the blind and visually impaired to connect with friends far away, without having to rely on friends immediately nearby (a predicament that a Facebook study found often deters the visually impaired from participating at all). Until then, there are things you can do as a user to make social media more inclusive to everybody on your friend list (this toolkit from the feds is a good place to start). In the meantime, don’t worry, Facebook can’t describe how you look in that swimsuit — yet.
Michael Andor Brodeur can be reached at firstname.lastname@example.org. Follow him on Twitter @MBrodeur.