Experimental design, data collection and analysis

Generative artificial intelligence (GAI) tools can support research design, methodological planning, data processing, and analytical workflows.

GAI tools may enhance productivity and reproducibility when used with appropriate oversight. These tools can be used to facilitate the methodology of data collection, but should not be used to conduct the data collection for you - researchers retain responsibility for methodological decisions and must not delegate core research judgement to AI systems. See UKRIO research integrity guidance.

Read some use cases below.

Experimental and study design

• Drafting/reformatting protocols
• Designing surveys or interview frameworks
• Designing data pipelines or database structures.

Data collection, processing and analysis

• Transcription and translation
• Generating ‘dummy’ data
• Data cleaning, reformatting, and quality control
• Code generation and troubleshooting
• Suggesting appropriate statistical tests
• Generating data visualisations.

Policy and governance considerations

• Data uploaded to public AI tools may create GDPR, confidentiality, or IP risks. Researchers must not input personal or sensitive data into tools that do not meet institutional data protection requirements. Read our research data management policy
• Use of AI in data analysis should be disclosed/documented to support reproducibility, transparency, and ethical conduct of research
• Compliance considerations should include discipline-specific methodological standards
• If AI tools form part of the research design or data processing workflow, this may need to be reflected in ethics applications. Participants should be appropriately informed where required by research ethics approval.