3 Ways AI Is Making The Internet More Accessible

3 Ways AI Is Making The Internet More Accessible




According to a study, 71% of disabled customers with access requirements click away from a website that they find difficult to use. Also,

Those customers who click away have an estimated spending power of £11.75 billion in the UK alone. This is approximately 10% of the total UK online spend in 2016.
82% of customers with access requirements will spend more if websites were more accessible.

Web developers design websites with a focus on delivering appealing visual components to satisfy their clients. However, they tend to overlook the need to include web accessibility components. Users with disabilities, therefore, shun disabled-unfriendly websites, thereby increasing website bounce rates.

Websites must be equally accessible to all users, and AI has already begun taking part in this important mission. Most business owners are unaware of website accessibility laws and that failure to comply with the Web Content Accessibility Guidelines (WCAG) will result in losing a major segment of potential customers as well as serious litigation. It should be noted that in 2018 alone, 2,258 website accessibility lawsuits were filed in the federal court.

Maintaining compliance with Web Content Accessibility Guidelines (WCAG) helps businesses build brand loyalty with their disabled clients. Accessible websites contribute to an augmenting customer base through their benefits of usability and search engine optimization.

According to the Pew Research Center, 50% of disabled Americans are less likely to use the internet daily when compared to 79% of users who are not disabled.

Despite stringent acts like the ADA, several websites remain non-complaint. Web developers must consider automated, AI-empowered web accessibility tools like accessiBe. This groundbreaking tool automatically applies accessibility standards to websites, making them accessible to users with disabilities.

The AI-based solution scans websites, incorporates the required modifications and transmits accessible content best suited for users with disabilities. They also re-scan and work daily, a feature advantageous for dynamically changing websites.

How AI Helps With Digital Inclusion

AI has enabled easy access to information for users who have sensory, physical or cognitive impairments. AI helps by enabling disabled users to proactively interact with the web in the following ways:

1. Improving Website Accessibility

Alt text empowered by Artificial Intelligence automatically writes captions for the blind and the visually impaired. This form of text helps users with visual disabilities understand an image and its components. Without the inclusion of AI, users can hear only the word “image” without a description of its details.

Facebook uses AI-enabled automatic alt text that helps blind or visually impaired users feel connected and included. According to Facebook, 2 billion photos are posted across social media platforms and with nearly 39 million blind users and 246 million users with severe visual impairment, this AI-enabled automatic alt text helps them to make the most of Facebook.

Automatic alt text is based on object recognition technology built on a convolution neural network that has been tuned to millions of learnable parameters. When a visually disabled user accesses the screen reader on an iOS device, he can listen to the list of items in the photo on the Facebook page when he swipes past it. Earlier, users were only able to hear the name of the person who shared the photo. With automatic alt text being based on AI, users obtain access to a richer description of the photo.

“Think about an eCommerce website with tons of banners that include text in the form of offers and promotions,” notes Shir Ekerling, Co-Founder & CEO of accessiBe. “Blind people completely miss that kind of text messaging if they’re not described manually word-by-word, image-by-image. However, it’s practically impossible to manually keep track of every image on a given website. The object recognition and O.C.R (optical character recognition) technologies we utilize help blind people use websites on a daily basis.”

The learning element of the neural network detects images and visual concepts. It splits the image into three broad groups of people, objects, and scenes. The alt text details the number of people and their actions followed by the number of objects in the image. It also describes the settings and scenes detected in the image.

2. Opening Newer Channels to Access the Web

Voice interfaces have become common, but they aren’t equally accessible. A user with a speech disorder due to a neurological impairment like ALS or multiple sclerosis may not be able to use Google Assistant or Amazon’s Alexa.

Google’s Project Euphonia can make speech technology more accessible to disabled users. It uses AI to improve the capability of computers to understand diverse speech patterns that include impaired speech. Voices of people are recorded and used to optimize AI algorithms so that they can reliably transcribe words. Google software converts the recorded voice samples into spectrograms that are used to train the AI algorithm to identify this less common form of speech.

This collaborative initiative by ALS Therapy Development Institute (ALS TDI) and Google will help users who cannot speak or have physical disabilities access the web independently by detecting gestures and sounds.

3. Boosting Customer Engagements

AI helps customers with disabilities to deeply engage through ways like:

  • Text Summarization

Web accessibility solutions needn’t be limited to speech, visual, and mobility concerns. Users with different grades of cognitive impairments or learning disorders also need help to access the web.

Text summarization is a tool that compresses a huge amount of information. It helps people with cognitive impairments such as dyslexia, attention deficit disorders, and memory issues.

Salesforce has developed an AI-based abstract summarization algorithm that condenses longer forms of text. This model uses deep learning neural networks to interpret content and introduce the right words to produce an accurate summary. This tool is based on the automated evaluation metric known as Recall-Oriented Understudy for Gisting Evaluation (ROUGE) that updates itself with higher scores for better summaries.

These AI-enabled text summaries are accurate and can help users with cognitive issues keep up with the deluge of information on the internet.

  • Lip-Reading

Providing closed captions can help expand the customer base for video and audio content.

The internet has a deficit of disabled-friendly content. Google’s DeepMind was developed based on AI’s lip-reading algorithm. This system can generate closed caption content for deaf users that helps the hearing impaired easily interpret visual content and improve web accessibility.

Google’s DeepMind was trained based on 5000 hours of six different television programs. This system successfully identified 17,500 unique words. It can outperform professional lip-readers by translating 46.8% of words without errors. The technology can enhance hearing aids, support silent dictation in public places, and facilitate speech recognition in a noisy environment.

Wrapping Up

The above-mentioned AI technologies support users with disabilities to easily access the internet for their needs. Developing accessible websites will help business owners attract and explore a newer segment of customers, drive traffic and increase sales.

Image Credit: AlisaRut/Shutterstock



Source link