University Students' Perceptions and Practices of Artificial IntelligenceApplications in Chemistry A Questionnaire Survey.xlsx

Abstract

This dataset contains de-identified questionnaire data from a cross-sectional survey on university students' perceptions and practices of artificial intelligence (AI) applications in chemistry. The survey was conducted online via the Wenjuanxing (Questionnaire Star) platform from September 6 to 9, 2026, targeting undergraduate and graduate students majoring in chemistry and related disciplines at Qinghai University, China. A total of 779 valid responses were obtained after data cleaning. The dataset is stored as a single Excel (.xlsx) file with 780 rows (1 header row + 779 responses) and 98 columns, comprising 7 metadata columns and 91 questionnaire data columns. The questionnaire is organized into six modules: 1. Demographic information (Q1–Q4): gender, education level, major category, and institution type 2. Current AI tool usage (Q5–Q9): usage frequency, application scenarios, overall satisfaction, perceived advantages, and common problems 3. Academic ethics and norm awareness (Q10–Q15): journal AI policy awareness, attitudes toward undisclosed AI-assisted writing, disclosure scope, views on AI co-authorship, institutional regulation status, and need for institutional norms 4. Critical verification and laboratory safety (Q16–Q25): AI error encounters, verification behaviors, perceived tool limitations, need for explainable AI, self-assessed critical evaluation ability, use of AI-generated experimental protocols, safety verification practices, encounters with dangerous protocols, agreement with experimental verification principle, and need for AI content in laboratory safety training 5. Educational needs and curriculum expectations (Q26–Q30): current curriculum AI content coverage, desired AI-chemistry courses, competencies needing improvement, views on mandatory AI use models, and core competencies in the AI era 6. Open-ended question (Q31): additional views and suggestions on AI development and application in chemistry Multiple-choice items are stored as binary columns (1 = selected, 0 = not selected); single-choice and five-point Likert scale items are stored as integer codes (1–5); the open-ended item is stored as de-identified text. Metadata includes serial number, submission time, completion duration (seconds), access source, source details, IP address with geographic location, and total score. All personally identifiable information (PII) such as names, student IDs, phone numbers, and email addresses was not collected during the survey. IP addresses are used solely for geographic distribution analysis and are not linked to individual identities. Open-ended responses were manually screened, and any potentially identifiable information was removed or generalized prior to deposition. This dataset can be used for secondary analysis of AI chemistry education status, investigation of factors influencing students' AI ethical perceptions and critical thinking, needs assessment for chemistry education curriculum reform, cross-regional comparative studies in AI education, and statistical methods teaching.