Studying the brain using functional MRI is a complex and effort-intensive process. Conventionally, expert human input is required at virtually all stages of the workflow, ranging from experimental design to preprocessing to interpretation of results. In recent years, however, a vibrant neuroinformatics ecosystem has developed that promises to substantially reduce the effort burden on investigators by automating many of the typical processing and analysis steps. In this talk, I review a number of new tools, platforms, and standards--including the BIDS standard, the Nipype framework, and the OpenNeuro, NeuroVault, Neurosynth, and NeuroScout platforms--that collectively offer to dramatically improve the efficiency and scale of fMRI data analysis. I argue that widespread adoption of these tools would substantially reduce human effort and investigator bias, thereby freeing up researchers to focus their energies primarily on those aspects of the typical workflow that are best handled by humans--namely, hypothesis generation, experimental design, and model specification.