Improving Accessibility of AI Technology Development for a Diversity-Inclusive Sustainable AI Society

2020 INNO-vation Disruptive Challenger

With AI technology becoming more prominent in recent years, large companies such as those involved in IT have been proactively working with AI technology designed to support those with disabilities. Due to Japan’s shortage of human resources brought about by the country’s low birth rate and aging population, AI technology that helps those with disabilities to play a larger role in society is becoming increasingly vital. Although the use of AI technology is increasing, those with visual impairments are almost entirely unable to partake in the development process. Even if new programs are developed from a software 2.0 point of view, the analytical methods to interpret data remain largely the same, leaving those with visual impairments out in the dark, a situation often referred to as informationally impaired. That being said, as mentioned in the Japanese government’s AI strategy, AI technology is meant to “promote a sustainable society that embraces diversity”, and that very diversity is something that should be carefully considered during the development process.

In recent years, there’s a tendency when using data-driven recursive technology such as Deep Learning that the created AI tech will exhibit a strong bias. Therefore, a diverse group of people should be involved in the development process. In order to realize such a society that promotes sustainability and embraces diversity, this challenge was focused on improving the accessibility of AI development for the visually impaired, as well as establishing a data science environment and creating the associated primer content (written below).

• Data Science Environment Improvement for the Visually Impaired Project
• Data Science Primer For Those Using Screen Readers