How can generative AI support your everyday research practice in the natural sciences without replacing critical judgement, subject expertise or scientific responsibility? This online workshop series is designed for PhD Candidates and Postdocs who want to develop a sound understanding of how generative AI works, where its limitations and risks lie, and how legal, organizational and academic guidelines shape its use in research contexts.

The first session focuses on basic principles, data protection, confidential research data, academic integrity, transparency, documentation and structured prompting with KI-Chat@JGU. Between the two sessions, participants will test AI applications in their own research or work context. The second session builds on these experiences and focuses on prompt refinement, quality assurance, context engineering and the development of responsible AI-supported workflows.

The central question throughout is how AI can support scientific work in a useful and accountable way while keeping scholarly responsibility and expert assessment with the researcher.

Language-proficiency level minimum English C1 (CEFR).

PhD Candidates and Postdocs of the Natural and Life Sciences, Mathematics, and Psychology.

Content:

• Basic principles of how generative AI works, including key limitations and risks such as hallucinations, bias, data flow and context dependency
• Legal, organizational and academic guidelines for responsible AI use, especially data protection, confidential research data, documentation, transparency and academic integrity
• Structured prompting with KI-Chat@JGU, using participants’ own research-related tasks and examples
• Reflection on participants’ trial experiences between the two sessions, including successful prompts, open questions and problematic AI-generated results
• Advanced prompting, quality assurance and context engineering for more complex natural science workflows, including information structuring, documentation, project organization and preparation for analysis

Learning objectives:

Participants will

• be able to assess opportunities, limitations and risks of generative AI in their own natural science research context
• be able to apply relevant guidelines and principles of responsible AI use to typical academic work situations
• be able to formulate, test and improve prompts for increasingly complex research-related tasks
• be able to evaluate AI-generated output critically and make it more transparent, controllable and scientifically accountable
• develop first ideas for responsible AI-supported workflows and reusable context structures in their own research practice