How generative AI is transforming agri-food systems research

Research objective The emergence of generative AI is expected to reshape agri-food systems research by changing not only individual research tasks but the entire lifecycle of knowledge production in interaction with its stakeholders. This transformational potential, combined with ethical concerns that may undermine research integrity, urges researchers to rethink their role: from producers of scientific knowledge to participants in a broader innovation ecosystem characterized by AI-mediated co-creation of hybrid intelligence. This paper examines how generative AI can responsibly advance agri-food systems research by transforming the foundations of research data, models, and knowledge infrastructures. Main findings In terms of research data, AI can enable the use of diverse, heterogeneous, and dynamically evolving data sources, supporting end-to-end automation of the data lifecycle, including acquisition, processing, and visualization. For research models, generative AI facilitates data-driven and hybrid approaches, enhancing the full model lifecycle from ideation and design to validation and deployment in agentic AI-systems. Regarding research knowledge, advances in large language models (LLMs) and natural language processing (NLP) enable more efficient capture, integration, interpretation, and dissemination of information. Retrieval-augmented generation (RAG) systems further exemplify these capabilities by combining LLMs with curated knowledge bases to improve contextual relevance and reduce hallucinations. At the same time, generative AI raises critical ethical challenges, including issues of scientific integrity, security, privacy, environmental impact, intellectual property, and digital sovereignty, highlighting the accountability of scientists to guide its development and application in accordance with principles of responsible AI.