Commentary: The rising absolute number of neuro-inspired AI papers represents a substantial footprint

In a recent analysis, Renato Duarte examined the intersection of neuroscience and artificial intelligence (AI) through the lens of citation patterns. He found that, while AI has been loosely inspired by established neuroscientific concepts rather than directly utilizing specific neural mechanisms, many design choices in current AI tools parallel properties of the brain. These include fast activity-based reasoning mechanisms, geometric codes representing relative positions, and planning through mental simulation.

To illustrate the growing connection between the two fields, Duarte's analysis compared the absolute number of AI papers citing neuroscience to the fraction of AI papers doing so post-2015. The plot revealed a steep increase in cross-field citations for both neuro-citing-AI and AI-citing-neuro papers around that time. However, the fraction of neuro-citing-AI papers increased, while the fraction of AI-citing-neuro papers remained low, possibly even decreasing.

So, is neuroscience having a growing or diminishing impact on AI? On one hand, the declining fraction of neuro citations suggests it may not be a dominant driver of the AI boom compared to other factors. On the other hand, the rising absolute number of neuro-inspired AI papers still represents a significant footprint, even if it constitutes a decreasing percentage of the overall AI literature.

For researchers working at the neuroscience and AI interface, this raises a core question in the latest post-2022 AI boom. Is AI's impact so vast that even a fringe possibility of influencing the field is a competitive strategy for maximal impact? Or is the AI field on a difficult-to-shift path, with lower-hanging fruit to be found in applying AI to understanding the brain? Regardless, Duarte calls for new results and tools to refresh the interdisciplinary relationship between these fields.