A Framework for Teaching the Responsible and Effective Use of Generative AI in Statistics and Data Science
Abstract
The rapid rise of generative artificial intelligence (AI) following the public release of tools such as ChatGPT has begun to reshape how students engage with coding, statistical analysis, and written communication in higher education. Growing use of conversational AI systems creates opportunities for learning alongside challenges around reliability, transparency, privacy, and appropriate academic and professional practice, highlighting the need for structured education in their responsible use. This paper presents a framework for teaching the responsible and effective use of generative AI in statistics and data science. The framework is organised around four areas: Understanding, Development, Reporting, and Safety, covering conceptual foundations, practical interaction within analytical workflows, documentation and transparency, and issues including reliability, bias, privacy, and ethics. Its design draws on emerging work in AI literacy and established principles from statistics and data science education, and was further refined through workshop and conference activities and participant feedback. The framework has been implemented across multiple settings, including a massive open online course, short training sessions, and undergraduate and postgraduate skills-based teaching. We use these experiences to identify broader design considerations for educators, including the value of authentic analytical tasks, critical evaluation of generated outputs, integrated consideration of responsible practice, and flexibility as generative AI technologies continue to evolve.
Related
- Guide to Teaching with Data in Statistics and Data Science Classrooms
- Two-way Guidance: Simultaneous Learning of Summary Statistics for Generative Bayes
- Effective atom-dimer bands of three-body lattice systems at strong coupling: exact critical mass ratios and a statistics-quasimomentum duality
- A New Framework for Interfacial Statistics: Exact n-Point Correlations of Gaussian Level Sets
- Generative Marketing Mix Modeling: A Causal Inference Framework Linking GEO and GEM to Business Impact
BibTeX
@misc{davies2026framework,
title = {{A Framework for Teaching the Responsible and Effective Use of Generative AI in Statistics and Data Science}},
author = {Vinny Davies and Craig Alexander and Jennifer Gaskell and Matthew Parker and Jake Lever},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.21663},
eprint = {2610.21663},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.21663},
}
RIS
TY - UNPB
TI - A Framework for Teaching the Responsible and Effective Use of Generative AI in Statistics and Data Science
AU - Vinny Davies
AU - Craig Alexander
AU - Jennifer Gaskell
AU - Matthew Parker
AU - Jake Lever
PY - 2026
DA - 2026/10/15
PB - arXiv
DO - 10.48550/arxiv.2610.21663
UR - https://arxiv.org/abs/2610.21663
AB - The rapid rise of generative artificial intelligence (AI) following the public release of tools such as ChatGPT has begun to reshape how students engage with coding, statistical analysis, and written communication in higher education. Growing use of conversational AI systems creates opportunities for learning alongside challenges around reliability, transparency, privacy, and appropriate academic and professional practice, highlighting the need for structured education in their responsible use. This paper presents a framework for teaching the responsible and effective use of generative AI in statistics and data science. The framework is organised around four areas: Understanding, Development, Reporting, and Safety, covering conceptual foundations, practical interaction within analytical workflows, documentation and transparency, and issues including reliability, bias, privacy, and ethics. Its design draws on emerging work in AI literacy and established principles from statistics and data science education, and was further refined through workshop and conference activities and participant feedback. The framework has been implemented across multiple settings, including a massive open online course, short training sessions, and undergraduate and postgraduate skills-based teaching. We use these experiences to identify broader design considerations for educators, including the value of authentic analytical tasks, critical evaluation of generated outputs, integrated consideration of responsible practice, and flexibility as generative AI technologies continue to evolve.
ER -