Over a four month period, we engaged with over 100 research software engineers at a number of community events (SciPy, RSECon 25, US-RSE 2025) to better understand the impacts of Generative AI (GenAI) on the work of Research Software Engineers (RSEs). Through these conversations we have seen an interest from the RSE community to make a public statement about how GenAI is impacting RSEs and the vision that RSEs have for their profession in this new technological landscape.
Research software engineering – building and maintaining software to enhance research – is an indispensable part of modern research. Since computing entered the research toolkit, RSEs have been integral to research across disciplines through the software and infrastructure they build and maintain.
The role of research software engineers spans a wide spectrum, from supporting individual labs to collaborating with campus-wide IT organizations. As universities move rapidly to integrate GenAI into their operations, RSEs are increasingly responsible not only for using these tools but also for incorporating them into institutional infrastructure.
With the recent rapid emergence of GenAI, the conversation around software engineering in the age of GenAI has too often been focused on AI hype and/or dystopian futures predicting the demise of software engineers. This has left relatively little space for nuanced discussions around this topic, and the complexities of developing research software in a future where GenAI is widely available.
Sustainable research software rises from collaboration. RSEs are the linchpin: they bridge disciplines, turn research needs into working systems, and mentor colleagues in sound computational practices. These human roles – translation, coordination, and trust-building – are not functions GenAI can adequately replace.
Like many tools RSEs use in their work, GenAI can be powerful if applied responsibly and appropriately. Current experience with GenAI suggests that these technologies are helpful in an assistive capacity, but they are often found lacking when applied to complex, multifaceted problems. RSEs and researchers should own technical decisions and be trusted to select and integrate technologies that best advance the organization’s mission.
Research is an iterative process that relies not only on findings published across generations, but also on the methods and infrastructure that produce them. As these methods and infrastructure grow more complex and computational, we must ensure that their development is reproducible in order to maximize research impact. The introduction of GenAI into the research process raises serious new questions about how it fits within best practices for reproducible research. Because many GenAI tools are opaque in their function, adoption should proceed cautiously and be accompanied by refreshed reproducibility standards. Decisions about the use of GenAI in research projects should be clearly documented, preferably decided in advance and incorporate the views of the RSEs involved.
The widespread emergence of GenAI is disrupting many sectors, including the RSE community. As with any new technology, time to experiment and guidance on when and how to use it are critical for integrating GenAI successfully into the research enterprise. Closing knowledge gaps depends on professional development, mentorship, and the sharing of new ideas – areas often hobbled, rather than helped by GenAI. RSEs need tailored training, best practices, and institutional policies that reflect the realities of research software development, not just generic software engineering. The community must also invest in mentoring and knowledge exchange so junior RSEs build strong foundations and grow into future RSE team leaders.
The change of capabilities, expectations and understanding of GenAI has been rapid and development is likely to continue apace for the foreseeable future. As Research Software Engineers working at the forefront of research and technology, we recognise the need to adapt, adjust and develop as the situation changes. Agility sits at the heart of software engineering and Research Software Engineers are best placed to steer their position as things develop.
As Research Software Engineers, we recognize that our work is critical to the success of modern research across all sectors. With the emergence of GenAI, RSEs are tasked with the challenge of evaluating and incorporating the technology into a complex ecosystem of research infrastructure. We must ensure that GenAI is used responsibly and in a way that is compatible with the ethos of the modern, open, computational research enterprise, and this can only be accomplished through careful evaluation, implementation, and training.