This workshop addressed challenges and opportunities surrounding access to online content and services, including rich, non-text content. Consumers are increasingly relying on online information for guidance on matters of health, education, and other important topics. Our ability to provide online access for consumers generally, including people with disabilities, must keep pace.

The workshop brought researchers and stakeholders from the disability community together with researchers from relevant areas of computing research (such as computer vision, or natural language processing) to identify strategic opportunities to solve pervasive accessibility challenges through computing.

The workshop included four parallel topical sessions, with unifying plenary presentations. Each topical session will include two participants who are people with disabilities, and/or disability advocates; six computer science researchers, including at least 3 who are not already active in research on disability and technology, and one representative of an interested federal agency.

Topics to be considered included:

• Automatic detection of document structure to make textual content accessible to blind and visually impaired users: As mentioned earlier, making textual content accessible to blind users requires training of all who create or prepare documents for presentation online. Research is needed
• Automatic description of image and video content for educational materials and other online materials: Adapting textbooks and other materials that include pictures or diagrams requires expensive manual annotation to describe these images for readers who cannot see them. New regulatory requirements (##21^st Century Communication and Video Accessibility Act) are increasing the need for audio description of video materials, for blind people; today this, too, is an expensive manual process. Research is needed on how to prepare such information automatically.
• Automatic simplification of content: Many people have difficulty comprehending complex ideas, or material written using an extensive vocabulary. For example, many people who need to access information about support programs offered by the Social Security Administration or other agencies are experiencing cognitive decline, and have difficulty finding and comprehending descriptions of programs and the conditions of eligibility. Today, efforts to meet these needs, when undertaken at all, require expensive manual work. Research is needed on ways to automatically transform information on line so as to make it easier to understand.
• Software architecture for configurability: The needs of people with disabilities are very diverse, so means are needed to permit people to specify how they want information and services to be presented to them, and how they wish to interact with them. Progress is being made to support personal, online specifications of these preferences (see GPII.net), but, today, a great deal of manual work is needed to create software than can respond to these varying needs. Research is needed on how software user interfaces can be automatically reconfigured as needed.
• Access to large quantitative datasets: In an era of increasingly open access to data, many federal agencies are making large amounts of quantitative data available for public use. Unfortunately these agencies lack efficient, effective means of making these data accessible to people who cannot see information visualizations.
• Automatic simplification of interactive web sites and applications: Access to services, as well as to content, needs to inclusive. How can we leverage machine learning techniques and data make the interactions that a software application requires simpler, for the benefit of users who often find online interactions confusing, including users with cognitive limitations?
• Inclusive design of games and simulations: These highly interactive online offerings are increasingly used for education, and for promoting public engagement. How can they be made accessible for learners with disabilities?
• Access to map and geographic information: Recent technical developments make it very easy to include maps in online content of all kinds, for a wide variety of purposes, including presentation of geographically referenced data as well as navigation. Yet we lack accepted and effective approaches to making this information work for people who can’t see well or cannot process complex visual information. How can we provide effective access to map data?

The workshop also included a brainstorming session in which a wider range of more speculative opportunities, suggested by the participants, will be identified.

The workshop will lead to a published white paper, describing an agenda of computer science research that can address the challenges and opportunities discussed.