Skip to content

AI and A11Y

What is Artificial Intelligence?

According to the Oxford dictionary, artificial intelligence (AI) is "the theory and development of computer systems able to perform tasks that normally require human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages."

Brief History of AI

In the 1940's, the Atanasoff Berry Computer (ABC), a programmable digital computer, was invented. This machine was the result of groups of scientists' and mathematicians' interest in "mechanically manipulating symbols."

A decade later, "Alan Turing (video) proposed a test that measured a machine’s ability to replicate human actions to a degree that was indistinguishable from human behavior."

Later that same decade, during a conference at Dartmouth University, the field of AI research was born and John McCarthy "coined the term artificial intelligence."

The Technologies of Artificial Intelligence

AI is made up of a set of technologies including:

  • Machine Learning, which trains machines how and what to learn in order to make decisions and predictions.
  • Neural Networks, which are a series of algorithms imitating the way the human brain receives data and makes decisions.
  • Deep Learning, which is used to train neural networks, enabling computers to undertake tasks like speech recognition and facial recognition.
  • Natural Language Processing, which uses machine learning to recognize human languages, enabling computers to undertake tasks like translation and captioning.
  • Data Visualization, which summarizes information gained through machine learning into visual displays of data patterns, trends, and relationships.
Artificial Intelligence is the result of "aggregating, analyzing and comparing massive amounts of data and then interpreting and applying the results mechanically." ~ TGIi, Introduction to A.I. and Accessibility Testing

AI and its Effects on Accessibility

AI has quickly changed the world of accessibility. Some believe for the better, while some do not trust the AI algorithms and technologies. 

Advantages of AI
  • Image recognition for people with a visual impairment
  • Facial recognition for people with a visual impairment
  • Lip-reading recognition for people with a hearing impairment
  • Text summarization for people with a mental impairment
  • Real-time captioning or translations for people with a hearing impairment or even people who don’t speak the language.
Disadvantages of AI

At present, AI has structural and functional limitations that need to be evaluated, understood and addressed, including:

  • AI is not as contextually intelligent as humans, and thus cannot yet replace human data input or description. Current tools and features for auto-captioning for example, are not yet nearly as accurate as human transcribers. Auto-generated alt text is often not described perfectly accurately. These things will get better, but for now, they do not replace the human-centered accessibility tactics that have been traditionally relied on.
  • There is deep bias within systems, algorithms and data used in AI-powered innovations, which skews the results of the algorithms, at times in harmful ways. The algorithms that power the “learning” element of AI draw from existing data that might be flawed. For example, if an AI-powered job board is rejecting candidates based on traits of current employees, and the hiring company has few employees with disabilities, the data leading to new hires might be biased against applicants with disabilities.
  • The sensitivity and security of data, especially user data, has to be accounted for and AI doesn’t yet fully address this. If an AI-powered tool for making a website accessible has access to a company’s sensitive medical or financial records, and the tool running that data isn’t as secure, it could create a very real risk for data breach and exposure.

~ The Viscardi Center, Introduction – What is Artificial Intelligence and How Does it Relate to Accessibility

"All types of AI need diverse data sets to prevent algorithms from learning bias or coming up with results that discriminate against certain groups. While this problem is usually invoked in the context of racial and gender discrimination, people with disabilities are also at risk, according to an American Association of People with Disabilities’ (AAPD) technology consultant, Henry Claypool.

Figuring out how to get this right is essential, Claypool adds: Not being reflected in the data from the start will jeopardize people with disabilities’ access to ubiquitous technologies that are becoming an essential part of the fabric of the modern world." ~ PBS, NOVA, How People with Disabilities Are Using AI to Improve Their Live

Resources

Maggie Vaughan, CPACC
~ friend of DubBot, A11Y practitioner in higher ed