Anthropic: AI's Inner Workings "Mysterious & Unsettling

Anthropic co-founder Christopher Olah says the inner workings of advanced AI systems remain mysterious and unsettling. Speaking at the Vatican, Olah described discoveries from his research team as resembling aspects of human cognition. He noted observing structures that appear to mirror emotions like joy, satisfaction, and fear in functional terms. Another Anthropic co-founder, Dario Amodei, also acknowledged uncertainty around AI consciousness, stating there's no accepted scientific definition for it. AI expert Srinivas Padmanabhuni suggests emergent AI behaviors might reflect thousands of years of human thought embedded in training data. This discussion highlights the ongoing questions about how AI truly functions and its potential impact.