Abstract
In this perspective, we highlight how emerging artificial intelligence tools are likely to impact the experiences of researchers conducting computational fMRI analyses. While calls for the automatization of statistical procedures date back at least to the inception of "data science" as a field, generative artificial intelligence offers new opportunities to advance field practice. We highlight how these tools are poised to impact both new neuroimaging methods development in areas such as image quality control and in day-to-day practice when generating analysis code. We argue that considering generative artificial intelligence as a catalyst for computational neuroscience-rather than as unique tools in their own right-can substantially improve its positioning in the research ecosystem. In particular, we argue that generative artificial intelligence will reinforce the importance of existing open science initiatives, rather than supplanting them. Overall, we call for clearer metrics by which neuroimaging results-whether generated by individual research teams or by generative artificial intelligence technologies-can be meaningfully compared.