Work on the Use of Intelligent Voice Assistants by Older Adults Wins ACM TOCHI Best Paper Award
The paper by iSchool PhD student Alisha Pradhan, iSchool assistant professor Dr. Amanda Lazar, and University of Washington’s Dr. Leah Findlater explores the potential for and challenges of this technology for older adults who use digital technology infrequently.
Intelligent voice assistants like Alexa and Siri have become ubiquitous — in our homes, cars, and pockets. Such technology offers new possibilities for people with sensory, physical, and cognitive disabilities by facilitating access to information through voice inputs and removing the barriers inherent in the physical input of text (e.g., typing, spelling, etc.). Research at the Trace R&D Center has also explored the potential for the use of intelligent voice assistants by older adults, particularly those who only use digital technology infrequently.
Trace investigator and assistant professor in the iSchool Dr. Amanda Lazar and PhD candidate Alisha Pradhan, along with Dr. Leah Findlater, professor of Human Centered Design and Engineering at the University of Washington, set out to understand how some older adults perceive of and use this technology in a study deploying Amazon Echo Dots in the homes of older adults. Their resulting paper, “Use of Intelligent Voice Assistants by Older Adults with Low Technology Use,” recently won the 2020 Best Paper Award by ACM Transactions on Computer-Human Interaction.
Through participant interviews and review of device usage logs, researchers found that these older adults used the devices consistently for accessing online information, with health-related information being the most commonly sought. This finding leads to questions to be addressed by future work on information credibility in a non-visual, voice-only interface (particularly with respect to health information).
While using voice assistants for reminders may seem like a natural application of this technology for this population, the study revealed specific challenges. Participants expressed concerns related to incorporating the use of the device into their own routines (e.g., having to remember to set the reminder), as well as distrust of technology and technology infrastructure (e.g., power outage, internet failure), making them less likely to rely on the device for this purpose. Other challenges posed by the technology included the need to remember specific keywords for commands, errors in speech recognition, and the device timing out while the user was attempting to use it. These findings point to the need to increase accessibility and trustworthiness of the devices for use of intelligent voice assistants as memory support. One possible approach the authors propose is linking the voice assistant to physical objects that are conceptually and metaphorically understood by users (e.g., paper calendar).
With respect to current and future work in this thread Pradhan explains, “We continue working in this direction by examining older adults’ use of emerging voice technologies and by including older adults in the design of these technologies. In our ongoing work we are examining three different directions: In the first thread of our work we examined how unexplored “actors” (including social and material actors) shape older adults’ use or non-use of emerging voice technologies. This informs our second thread of work, where we account for these “actors” in designing personalized voice-based reminder systems to support aging. Finally, the third thread of our ongoing work examines how to include older adults directly in the design of voice technologies. Here, we co-design conversational voice experiences with older adults’ (e.g., by co-designing personas for voice assistants, involving older adults in writing dialogues for a customized voice agent).”
For more details, read the paper in ACM TOCHI and watch Pradhan’s presentation of this work at CSCW ’20: ACM Conference on Computer-Supported Cooperative Work and Social Computing in October, 2020.