Accessibility on social networks is an actual problem

The platforms have developed a variety of tools that can be used by everyone, including the blind, visually impaired, deaf and hard of hearing. It is clear that they still have a long way to go in this area.

Facebook, Twitter, Instagram, TikTok, Snapchat… Social networks have billions of users, but using these platforms is not necessary for everyone. Allowing individuals to connect, communicate and share, and yet some people may feel excluded, including those with disabilities. By this we mean the blind, visually impaired and the deaf and hard of hearing, whose access to social networks is not optimal.

Aware of this problem, the platforms have implemented functions and tools to make their services usable for almost everyone. Recently, Instagram announced the arrival of automatically generated subtitles for all videos on its app. Other tools have also been developed and implemented, but are they really effective?

Alternate text, option for improvement

One of the solutions available for blind or partially sighted people is alternative text. It is a written description of an image that can be accessed using a screen reader that reads text aloud using voice synthesis or displays it on a Braille display. Depending on the platform, this can be done automatically or manually. On Twitter, for example, users have the ability to write it themselves. Once published, this alternate text can be seen by hovering over the “ALT” mark in the lower left corner of the image.

Facebook, for its part, has relied on artificial intelligence (AI) for its tool. In 2016, it introduced a technology called Automatic Alternate Text (AAT) that uses object recognition to generate photo descriptions. In particular, a blind or partially sighted person using a screen reader hears what an image may contain. Facebook technology is able to recognize places, objects, activities and species of animals. He can also identify each person in the photo, but that no longer makes him a Meta measure published last November. The American giant has stopped using face recognition on Facebook, which automatically identifies people present in photos or videos posted on the social network. In other words, the AAT recognizes the number of individuals in the image, but no longer includes their names.

Whether manual or automatic, alternative text on social media still needs to be improved. Some platforms have limitations that make it difficult to add an appropriate description. By May 2020, Twitter users had only 420 characters to describe the image, forcing them to write a short description of the image. This limit has since been extended to 1,000 characters. And the feature on Facebook may be wrong or may not offer a complete description. Aware of the risk of mistakes, the social network also starts every description with “Maybe”. It also allows the user to modify automatically generated text.

Accessibility and entertainment option

Text-to-speech is another technology used on social media to help the blind and visually impaired. Allows automatically generated voice to read text aloud. Such features are available on TikTok and for Reels on Instagram. While these tools are useful, they can be problematic, even outside of these platforms. They can mispronounce words such as names or homographs. This risk is present in several languages, including French. A speech synthesis tool, for example, could be mistaken for the word “proud” which serves as an adjective and verb. Another obstacle to voice synthesis: language. Available on TikTok since December 2020, this tool is limited to several languages ​​and is not yet available in France.

Keep in mind that even if voice synthesis is created to help blind and partially sighted people, it’s also a way to make content fun. Social networks do not necessarily have to introduce such a feature to make their platform accessible. When Instagram announced its arrival at Reels in November 2021, there was no mention of accessibility. The platform launched it for its video creators to allow them to add narration without using their voice or humor.

Subtitles, a useful feature for the deaf and hard of hearing

To make their platforms accessible to deaf and hard of hearing people, several social networks like YouTube, Facebook or Twitter offer subtitles in their videos. They are generated automatically, and the platform interprets what is said and transcribes it. People can then play the content of the video. However, this is not always appropriate, as certain words can be misinterpreted, and social networks themselves are aware of this problem.

YouTube, which uses voice recognition technology to automatically create subtitles, says their quality can vary: “Automatic subtitles may not accurately reflect audio content due to pronunciation errors, accents, dialect use, or background noise. » It is for this reason that the video platform, along with others, allows users to check and edit subtitles. However, this option is not available on some social networks. Twitter, which has caught up with others by introducing automatic subtitles in late 2021, is yet to give its users the ability to change them. Another easier way is still to avoid mistakes: manually write subtitles to add them to the video.

Necessary development in the future

Among all these tools to help people with disabilities, the social network behaves like a bad student because it does not offer any: Snapchat. It only recently introduced a filter that allows users to learn American sign language, and has not announced any plans to make its platform more accessible. Either way, social networks still need to make progress in the area of ​​accessibility. An issue that also arises in relation to the metaverse, the universe of more immersive virtual worlds and which is considered by some to be the future of these platforms.

Leave a Comment