Imagining Home Entertainment AI Concepts
Funded by and cooperated with LG Electronics
My role: Project Manager & UX Researcher
Challenge: As applying AI into households is becoming prevalent, much has been discussed regarding home appliances (e.g. refrigerator, oven, vacuum cleaner). But what about home entertainment systems such as TV? What kind of innovative home entertainment systems could be designed and developed using AI?
Methodology: We conducted a series of systematic literature review to understand the concept and idea of 'home' and different AI functionalities that can be applied to the household context. We also held a co-creative design workshop with subject matter experts (e.g. HCI experts, designers, developers, and home owners).
Result: We created 5 different concepts and use-case scenarios of applying AI to innovative home entertainment systems.
Investigating Premium Oven Voice User Interface (VUI) User Needs
Funded by and cooperated with Samsung Electronics
My role: UX Researcher
Challenge: Ovens have been largely overlooked in terms of applying AI into home appliances. This is because ovens are perceived to be a straightforward system with few buttons and controls. However, what could be the needs of users if AI could be incorporated into ovens?
Methodology: We iteratively developed a voice assistant prototype that users can interact with during a series of Wizard-of-OZ studies. The prototype ranged from low-fidelity to high-fidelity to uncover user needs in different stages. The WoZ studies were conducted in both lab setting and in-situ home settings of participants. The collected data was analyzed by thematic analysis and classification of users' intents.
Big Screen TV Audio Satisfaction Improvement
Funded by and cooperated with Samsung Electronics
My role: UX Researcher
Challenge: It is difficult to measure and understand users' "true" satisfaction level of TV audio. When are they satisfied and dissatisfied? What factors (e.g. genre of the program, preferences of different household members, the number of people viewing the TV) play into their satisfaction level?
Methodology: We installed Raspberry Pi to study participants' TV and remote controls to log data such as TV channel, program, and volume level. We also sent messages through a chatbot we designed to ask their intents of volume changes, TV viewing context, and at-the-moment satisfaction level. At the end of the study, we conducted interviews to ask in-depth questions about the TV watching behavior, their perception of TV audio, and attitudes towards the chatbot.
Analysis: Through the collected data, we structured the behavior model of participants, analyzed the large scale of log data through SPSS using classification techniques, and analyzed interview data through thematic analysis.
Designing Chatbots for Black Americans with Chronic Conditions during COVID-19
My role: Research Lead
Challenge: Although COVID-19 has disproportionately impacted the Black American population, there has been lack of focus on how technology can support the needs of the group, especially those with chronic conditions. As chatbots have the potential to meet the needs of the Black American population due to their high accessibility and personalization, we wanted to understand what their needs were towards chatbot regarding protecting their health from COVID-19.
Methodology: We conducted interviews to understand the information needs and behaviors of Black American participants. We investigated their challenges in protecting their health and ways they have used technology to search for COVID-19 information. We also conducted design studies to understand their needs towards chatbots that support their health. We had participants design an imaginary conversation between themselves and their ideal chatbot to probe their needs in a realm of unlimited possibility.