Determine whether and how AI advances can contribute to cognitive science

Determine whether and how recent advances in artificial intelligence can contribute to addressing the challenges of cognitive science.

Background

The paper reviews rapid progress in AI, including task-performing models across vision, reinforcement learning, and language, alongside debates about their explanatory value and implications for theory building in cognitive science. Despite representational alignments between AI systems and brain measurements, the authors note ongoing disputes about model testing and interpretation.

Within this context, they explicitly state that it remains unresolved whether and in what ways AI progress can help address core challenges faced by cognitive science, motivating the agenda of naturalistic computational cognitive science laid out in the remainder of the paper.