Machine Learning & How To Make Your Website Accessible For The Disabled

According to the World Health Organization, upwards of one billion people worldwide experience some form of disability. This is about 15% of the world’s population. Learn how to make a website accessible for the disabled and the effect machine learning has on this.

In the U.S., the Centers for Disease Control and Prevention reports that 61 million adults, or approximately 26% of the country’s adult population, live with a disability. This 26% breaks down as follows:

• 13.7% have mobility issues
• 10.8% have cognition issues
• 6.8% have independent living issues
• 5.9% have hearing issues
• 4.6% have vision issues
• 3.7% have self-care issues

Check Out Our ADA Compliance For Websites Guide

What Is Web Accessibility?

The term web accessibility means that the World Wide Web, popularly known as the Web, and its plethora of websites should be available to everyone, including those with disabilities.

All people, including the disabled, need to be able to perceive, understand, navigate, interact with and contribute to the Web.

The U.S., Title III of the Americans with Disabilities Act (ADA) mandates website accessibility, as do various state and local laws. Both the World Wide Web Consortium (W3C) Web Accessibility Initiative and the Web Content Accessibility Guidelines (WCAG) set forth accessibility guidelines with which websites must theoretically comply or be subject to lawsuits.

Current Accessibility Failings

Unfortunately, a recent analysis of 10 million websites worldwide revealed the following:

• 98% failed to satisfy menu requirements
• 89% failed to satisfy popup requirements
• 83% failed to satisfy button requirements
• 71% failed to satisfy form accessibility requirements
• 52% failed to satisfy alt attribute requirements on images
• 22% failed to satisfy link requirements

Perhaps not surprisingly, the number of U.S. website accessibility lawsuits has dramatically increased in recent years.

In 2018 alone, 2,285 such lawsuits were filed. This number increased by 51.7% in the first half of 2019, the last period for which data is available.

Specific Accessibility Issues

The American Community Survey of 2017 found that only 21.6% of people with disabilities access the web, compared to 42.1% who find it impossible to do so.

If you or someone that you know has a disability, you understand the struggles of accessing the web. The following are issues when developers and website users fail to make their website accessible:

• If you’re vision impaired or blind, you can’t read screens.
• If you’re hearing impaired or deaf, you can’t hear videos.
• If you’ve lost an arm or hand to amputation, you can’t easily type on a keyboard or use a mouse or trackball to point and click.
• If you’re quadriplegic, you may have insufficient arm movement or finger dexterity to use a keyboard, mouse, or trackball.
• If you have epilepsy, flashing website content may trigger a seizure.
• If you have dyslexia, reading website text is difficult, at best.
• If you have a learning disability, website content can be confusing or hard to understand.
• If you are color blind, differentiating text from background colors or reading a graph, especially a pie chart

How To Make A Website Accessible For The Disabled

Artificial Intelligence and Machine Learning to the Rescue

Artificial Intelligence (AI) is the science of programming computer systems to mimic human abilities. Machine Learning (ML) is the subset of AI that teaches the machines how to learn so they can do this without having to rely on explicit human instructions or commands.

ML uses sophisticated algorithms and statistical models that let the machines analyze and draw inferences from various data patterns. Both AI and ML promise to create the tools necessary to allow people with a disability to fully access the Web. Companies can use an annotation tool to add information to a text, image, or other types of the database they have which will later help the AI model learn more efficiently.

Website designers and developers are tasked with an awesome responsibility: understand how to make a website accessible for the disabled and execute.

It is part of their jobs to make websites perceivable, operable, understandable, and robust, often by using AI-based tools. This includes the following:

• Providing keyboard navigation via the Tab button to accommodate people with motor impairments
• Providing audio content descriptions to accommodate people with visual impairments
• Providing optimum graphical elements, such as fonts, colors, and adequate spacing, to accommodate people with visual impairments
• Providing a built-in library that contains commonly used idioms, phrases, and slang to accommodate people with cognitive impairments
• Providing closed captioning on videos to accommodate people with hearing impairments
• Providing a facial recognition alternative that allows disabled site visitors to access password-protected pages without having to type in a password or enter a CAPTCHA code
• Providing automated tools that suggest ways in which disabled site visitors can correct minor problems, such as typos
• Providing robust content that assistive devices can reliably translate for disabled users
• Using H1 titles, H2 headings, and H3/H4 subheadings as necessary to make it easier for screen readers and other apps to easily organize the information
• Using tables only for tabular data to prevent confusion
• Allowing their sites to resize text in various browsers without breaking the site’s content; i.e., not using absolute units of measurement
• Not using media files, such as music or video files that automatically load and play when pages load

Existing Apps And Assistive Devices

Many AI- and ML-based apps and assistive devices exist and are in use today – and not “just” by the disabled. Here are a few examples

• Google’s Siri: a voice-controlled virtual assistant for iPhones. Users can use it to not only verbally dictate a text message to someone, but also to read the reply and do Google searches.
• Cortana: Microsoft’s answer to Siri. Cortana is a voice-controlled virtual assistant implemented on Windows.
• Amazon’s Alexa: A virtual assistant capable of voice interaction, streaming podcasts, and reading a website’s news, sports, traffic, and other information.
• Apple’s VoiceOver: A screen reader built into the macOS, iOS, tvOS, watchOS, and iPod operating systems. Users can access their Macintosh or iOS device utilizing spoken descriptions and, in the case of the Mac, the keyboard. Not only does VoiceOver assist the blind and visually impaired, but it also aids people with dyslexia.
• Chatbots: These are found on many websites today, simulate and process human conversation, both verbal and written. For instance, when you’re visiting a website and a little screen or box pops up asking if you need help or want to talk to a representative of the company whose site it is, this is a chatbot in action. Another example is when you order something online and a screen pops up, or you get an email, confirming your order and letting you know how much it will cost and when you can pick it up. Chatbots are taught to understand natural language and the contexts in which it occurs. They consequently enhance the user experience for all site visitors, whether or not they’re disabled.
• AccessiBe’s Automated Web Accessibility Tool: This tool runs in the background of a website, constantly analyzing its structure and components, such as dropdowns, pop-ups, forms, and internal windows. It generates keyboard navigation, including the use of a virtual keyboard. It also allows epileptic site visitors to disable animations and flashing GIFs that might otherwise cause them to have a seizure. Finally, it allows visually impaired site visitors to make needed adjustments to the site’s color contrasts, fonts, cursors, text sizes, spacing, and typography.
• Voiceitt’s Electronic Translator App: This app can be taught to recognize the user’s voice and the way he or she pronounces words. Once installed on the user’s phone or tablet, he or she speaks various words to train the app to recognize his or her voice and pronunciations and ultimately speak for them. Thus, speech impaired people gain the ability to be understood by others, including websites.

Conclusion

As AI and ML continue to evolve, there’s almost no end to the ways in which they can provide full Web accessibility to people living with a disability. When thinking about how to make a website accessible for the disabled, talk to your developers or download a plugin that will help make your website more accessible.