User Perceptions of YouTube's Recommender AI
Course: UX Design
Tags: Research, personas, semi-structures interviews, affinity diagramming, observations, surveys
Objective: Examine the YouTube recommendation feature and understand user practices and sentiments surrounding the recommendation system, with a focus on transparency, privacy, and user control. The project aimed to gain insights into how users interact with the YouTube recommendation feature and identify areas for improvement.
User research
Observations: utilized 'think aloud' and 'cognitive walkthrough' methods to observe user behavior and interaction flow.
Survey: Collected both qualitative and quantitative data on user preferences, satisfaction levels, and pain points related to the recommendation system.
Results
Usage of YouTube: Discovered diverse user practices, including daily and weekly usage for education, entertainment, and politics.
Satisfaction Levels: Found high satisfaction with recommended videos, with users valuing the feature for exploration and content discovery.
Uncovered pain points related to already seen videos, surveillance concerns, unwanted content, and privacy.
Analysis and opportunity areas
Employed Affinity Diagramming and User Personas to synthesize and define discovered practices.
User Personas and HMW Questions: developed personas ('Sceptical Simon' and 'Positive Pamela') and formulated How Might We questions addressing transparency and user control in data usage.
Concludes with a design proposal focusing on enhancing transparency, privacy, and user control within the YouTube recommendation system to improve the overall user experience.
