OK, who else is excited about the potential of large language models (LLMs) x Accessibility? 🙋♀️
I believe AI had the opportunity to significantly improve ease of access and experiences for people with disabilities IF done right. Given my Dyslexia, I use Bing Chat daily to summarize, do research, modify tone of things I write...it saves me hours of otherwise wasted time. Plus I can use voice dictation for the many days my wrists and fingers experience RSI.
But LLMs are not perfect, and there are instances of albeism. Ableism is a term that describes discrimination or prejudice against individuals with disabilities. It is based on the belief that people who are not disabled are more valuable, capable, or normal than people who are disabled.
In LLMs examples of albeism are:
· Unfair discrimination and material harm by perpetuating stereotypes and social biases, i.e. harmful associations of specific traits with social identities
· Exclusionary norms by reinforcing narrow or rigid categories that do not reflect the diversity and complexity of human identities.
· Toxic language that can incite hate or violence or cause offense. For instance, a LM might generate slurs, insults, or jokes that target people with disabilities or mock their conditions.
· Reproduction or amplification of ableist language, such as using disability as a metaphor for something negative or undesirable.