Safety2d ago
Safety
Cognitive Neuroscience Approach To AI Alignment
A researcher proposes using computational cognitive neuroscience to predict the mechanistic failure modes of transformative AI. The goal is to identify efficient interventions that remain viable even under rushed development timelines. This framework focuses on brainlike AGI to anticipate alignment risks. It offers a niche, theoretical perspective rather than a scalable technical solution.