Abstract for Thesis

I tried to collect all my thoughts into an abstract in preparation for the thesis. It’s to long but parts of it will flow into other areas.

This research proposal is about exploring the potential of voice assistant technologies for the blind as well people with visual impairments, the extent to which this technologies can provide and assist access to the world, and how their design could be improved to serve the specific needs of the target demographic.

Voice assistants, like Apple’s Siri or Amazon’s Echo devices, have established themselves over the last ten years within applications of entertainment and home automation. They are designed with a general public in mind, a one kind serves all approach. Despite not being designed specific as assistive technologies, people with visual impairments nonetheless adopt voice assistants at a steady growing pace. Voice assistants point to a depart from more traditional assistive technologies for people with visual impairments, like screen readers or voice recognition software. A circumstance which is also mirrored in the specific development environments of the main competitors in the voice assistant market. Amazon, Microsoft and others do have disability labs but their research results find little space in the final products.

As a rather young and complex technology, voice assistants enjoy being intensely studied right now. Particular attention has so far been paid to the algorithms and artificial intelligences underlying the technology and their biases (Bajorek, 2019; Danielescu, 2020), persona design (Danielescu, 2020; Humphry & Chesher, 2020), issues around privacy (Easwara Moorthy & Vu, 2014; Pal et al., 2020), the post-phenomenology of such assistants (Hector & Hrncal, 2020; Wiltse, 2020) and, on a smaller scale, voice assistants as assistive devices for people with impairments (Masina et al., 2020; Pradhan et al., 2018).

In the studies cited, voice assistants are primarily approached from a technical, social or legal perspective. Within the give case of voice assistants being adopted as assistive devices, additional problems arise from the perspective of ontological design. Contrary to applications in the entertainment sector, voice assistants as assistive technologies become existential technologies and fundamentally change they way people with visual impairments have access to the world. A circumstance that can be observed in such mundane practices as online shopping.

It is thus crucial to ask: What makes contemporary voice assistants attractive to people with visual impairments? How can we ensure, that these technologies are up to the task of being of existential importance? What can we learn from the design of traditional but comparative assistive technologies, like screen readers and voice recognition software?

The proposed research project is setup as a comparative study that extends into a prototyping phase. The general approach is human-centered design with a heavy leaning on ontological design, which is to say that in terms of data collection the respective lifeworlds of the participants, the social situation and actants of a given environment are integrated (opposed to a focus on functionality, with research done in labs). In a first round, ethnographic methods are applied to study and compare traditional assistive technologies for people with visual impairments versus voice assistants for the same set of tasks, like information retrieval, online shopping and interaction with computational systems.

The data collected serves as a baseline for workshops on prototyping interaction as well as situations. The underlying framework for this phase is based on an universal design approach, which “is the design and composition of an environment so that it can be accessed, understood and used to the greatest extent possible by all people regardless of their age, size, ability or disability”1. The prototypes will explore directions in the design of voice assistants as assistive devices by applying learnings from traditional assistive technologies. The workshops are also exploring problems that arise from the unmitigated application of voice assistants as assistive technologies in order to analyze the inherent risks in such practices.

The findings will be disseminated in lectures, peer-reviewed journals, workshops as well as an open-source published guide for designing voice assistants for people with visual impairments. The guide is produced for interested parties, especially for designers, developers and people or institutions that are actively working with the blind or people with visual impairments. Next to the research results, the guide will include tutorials and technological fragments, like code examples and device blueprints and as such, can be built upon in practice as well as in further research projects.