[Paper Review] Multitasking with Alexa Multitasking with Alexa: How Using Intelligent Personal Assistants Impacts Language-based Primary Task Performance

Justin Edwards, He Liu+5 more|arXiv (Cornell University)|Jul 3, 2019

Personal Information Management and User Behavior46 references4 citations

TL;DR

This study investigates how using intelligent personal assistants (IPAs) like Amazon Alexa affects language-based primary tasks, finding that cognitively demanding tasks involving language generation are significantly more disrupted than simpler copying tasks. Using a dual-task experiment, the researchers demonstrate that shared cognitive resources—particularly in working memory and language processing—make IPA interruptions more detrimental during complex writing tasks.

ABSTRACT

Intelligent personal assistants (IPAs) are supposed to help us multitask. Yet the impact of IPA use on multitasking is not clearly quantified, particularly in situations where primary tasks are also language based. Using a dual task paradigm, our study observes how IPA interactions impact two different types of writing primary tasks; copying and generating content. We found writing tasks that involve content generation, which are more cognitively demanding and share more of the resources needed for IPA use, are significantly more disrupted by IPA interaction than less demanding tasks such as copying content. We discuss how theories of cognitive resources, including multiple resource theory and working memory, explain these results. We also outline the need for future work how interruption length and relevance may impact primary task performance as well as the need to identify effects of interruption timing in user and IPA led interruptions.

Motivation & Objective

• To investigate the impact of IPA interactions on primary language-based tasks, particularly in multitasking scenarios.
• To examine whether cognitive load differences between writing tasks (copying vs. content generation) affect susceptibility to IPA interruptions.
• To explore how theories of cognitive resources, such as multiple resource theory and working memory, explain performance disruptions during IPA use.
• To identify design implications for IPAs that minimize negative impacts on user performance during concurrent language tasks.

Proposed method

• Conducted a controlled dual-task experiment with participants performing either copying or content generation writing tasks while interacting with an IPA (e.g., Alexa).
• Used a randomized interruption paradigm where IPA interactions were introduced at varying times during the primary task.
• Measured performance through task completion time, accuracy, and error rates for both copying and content generation tasks.
• Applied cognitive theories such as multiple resource theory and working memory load to interpret the results.
• Collected data using a within-subjects design to compare performance under interruption vs. non-interruption conditions.
• Controlled for task duration, interruption length, and relevance to isolate the effect of cognitive resource overlap.

Experimental results

Research questions

• RQ1 How does IPA interaction affect performance on language-based primary tasks that differ in cognitive demand?
• RQ2 To what extent do shared cognitive resources between IPA use and primary tasks (especially language generation) impact task performance?
• RQ3 How do interruption timing and relevance influence the degree of disruption during multitasking with IPAs?
• RQ4 In what ways do theories of cognitive load and working memory explain the observed performance differences?

Key findings

• Content generation tasks, which are more cognitively demanding and involve active language production, showed significantly greater performance degradation when interrupted by IPA interactions.
• Copying tasks, which are less cognitively demanding and involve minimal language generation, were less disrupted by IPA interruptions.
• The study confirms that shared cognitive resources—particularly in working memory and language processing—lead to greater interference during multitasking with IPAs.
• IPA interruptions were more disruptive when the primary task required high levels of language generation, supporting the hypothesis that resource overlap increases interference.
• The findings align with multiple resource theory, showing that tasks competing for the same mental resources (e.g., verbal processing) experience higher disruption.
• The results suggest that IPAs may unintentionally impair rather than assist users during complex language tasks, especially when users are already engaged in cognitively heavy writing.