The National Association of State Chief Information Officers (NASCIO) has published a new report that discusses recent trends and research in the use of generative artificial intelligence (GenAI) technology aimed at helping states improve disability empowerment efforts, and offers recommendations for state technology organizations pursuing those applications.
“The emergence of GenAI has created a unique opportunity for state technology leaders to improve accessibility for disabled people in the workforce and their state populations,” NASCIO said in releasing the use cases on July 9.
NASCIO cited research that shows 13 percent of the U.S. population had at least one disability as of 2023, and that the number of disabled people in the U.S. workforce totaled 22.5 percent.
“With both trends expected to continue rising, generative artificial intelligence (GenAI) has emerged as a tool that can greatly improve accessibility in the professional and personal lives of those with disabilities,” the group said.
“This is particularly relevant for state chief information officers (CIOs) as the United States Department of Justice recently released a final rule detailing the guidelines state and local governments must follow to ensure accessible webpages and mobile apps,” NASCIO said.
The NASCIO report also cited research from the University of Washington that pointed to GenAI’s usefulness in addressing the needs of people with disabilities – including those with cognitive issues and chronic illnesses – and said the findings point to “mixed results” currently for harnessing GenAI tech to meet those needs.
The report offers several recommendations to state tech officials who are exploring GenAI to help people with disabilities, the first of which is to engage all stakeholders when evaluating the use of AI tools.
The report further recommends the cultivation of inclusive data sets to combat AI-driven bias.
“Some AI models are built on datasets that inaccurately reflect or completely exclude certain demographics of a population, including disabled people,” NASCIO said. “GenAI tools built on non-inclusive datasets may favor non-disabled individuals, leading to biased outcomes. Inaccurate data labeling or excluding certain demographics from a dataset can perpetuate this bias.”
NASCIO also recommended embracing transparency in the development of AI tools including in the areas of data sources, and modeling and decision-making processes.
“Building confidence in GenAI is hampered by a lack of transparency in development,” the group said. “Disclosing this information allows end-users to hold companies accountable for AI shortcomings, fosters trust, and improves relationships and collaborative efforts between stakeholder groups.”
“The power of GenAI has yet to be fully realized, but people with disabilities are harnessing it for better, more accessible lives,” NASCIO said, adding, “current research analyzing use cases of available GenAI tools increasing accessibility is promising.”
