Perceptions and Utilization of GenAI Tools among Data Science Students and Faculty

Students use GenAI extensively but lack confidence in evaluating outputs: While 119 surveyed students reported substantial use of ChatGPT for coding and writing tasks, the study found that "confidence in interpreting AI-generated outputs was limited," indicating a critical gap between adoption and understanding that could lead to unchecked errors in data science work.

Faculty are underprepared to teach responsible AI use: Only 14 faculty respondents showed "limited" self-rated proficiency and infrequent classroom integration of GenAI, despite recognizing its importance—suggesting educators themselves need training before they can guide students on responsible AI practices.

Awareness of AI limitations increases with academic level: The research found that "familiarity with GenAI and awareness of its limitations varied more by academic level than by gender," meaning advanced students better understood risks like accuracy and reliability concerns, while earlier-stage students may be more vulnerable to over-reliance.

Structured guidance is urgently needed: The study explicitly calls for "structured training, validation practices, and clearer institutional guidance for responsible AI integration," highlighting that current ad-hoc adoption without formal frameworks leaves both students and faculty navigating GenAI risks without adequate support.

Common concerns about accuracy and reliability persist despite positive views: Even as students and faculty held "generally positive" perceptions of GenAI in data science, "concerns about accuracy, reliability, and over-reliance were common," suggesting skepticism exists but lacks actionable solutions.