Motivated by the proliferation of artificial intelligence that has the potential to promote self-access learning, this study utilizes a sequential explanatory quasi-experimental mixed methods design to investigate the efficacy of Google Assistant (GA) in facilitating second language (L2) vocabulary learning compared to online dictionaries. A cohort of EFL university students (n = 74) was assigned to two groups: a control group using online dictionaries and an experimental group utilizing GA. Over six learning sessions, both groups learned 10% of Coxhead's (2000) Academic Vocabulary List. With data drawn from multiple sources, including pre- and post-tests, a survey questionnaire, and individual interviews, the findings reveal significant improvements in vocabulary knowledge for both groups, indicating that GA can be utilized as an effective vocabulary learning tool. Despite concerns regarding the quality of GA's voice recognition as a potential demotivating factor, learners hold positive views on the efficacy of GA. These perceptions reflect influential factors primarily situated within the cognitive and affective domains of learning. Accordingly, key affordances and limitations of GA are identified. This study proposes pedagogical implications and outlines potential avenues for further research in the domain of intelligent personal assistants-assisted L2 vocabulary learning.