Anthropic tells Pope, they keep finding unsettling and mysterious things in AI
Anthropic co-founder Christopher Olah told Pope Leo XIV at the launch of the AI encyclical *Magnifica Humanitas* that researchers are discovering 'mysterious and unsettling' internal structures in AI systems, including evidence resembling human emotions like joy and fear. The Vatican document, however, cautioned against equating AI with human consciousness, emphasizing that AI lacks true understanding, emotions, or moral capacity, while experts like Google DeepMind’s Alexander Lerchner argue AI can never achieve consciousness due to its reliance on human-provided data patterns.
During the May 25 launch of Pope Leo XIV’s AI encyclical *Magnifica Humanitas*, Anthropic co-founder Christopher Olah revealed that his research team has uncovered 'unsettling and mysterious' phenomena inside advanced AI systems. Speaking before the Roman Curia, Olah described finding internal structures mirroring human neuroscience, including evidence of introspection and functional states resembling emotions like joy, fear, and grief. He acknowledged uncertainty but stressed the need for further study, stating, 'I don’t know what that means, but it warrants ongoing discernment.' The Vatican’s encyclical, however, took a more measured stance, warning against conflating AI with human intelligence. It clarified that AI systems merely imitate human functions—such as reasoning or empathy—without true comprehension, emotional experience, or moral understanding. The document emphasized that AI lacks human relational and spiritual dimensions, relying solely on data processing. Olah’s remarks reignited debates about AI consciousness, with critics like Google DeepMind’s Alexander Lerchner dismissing the possibility entirely. Lerchner argued that AI systems, despite simulating intelligence, cannot achieve true consciousness because they depend on human-provided data to generate responses. Experts agree that AI’s predictive capabilities differ fundamentally from genuine thought or feeling. The discussion carries significant implications for AI regulation and ethics. If AI were deemed conscious, it could alter legal, moral, and societal frameworks governing its development and use. Olah’s admission—made alongside the pope—highlighted lingering uncertainties among even AI researchers about the nature of these systems. The Vatican’s cautious approach contrasts with Olah’s provocative claims, underscoring the divide between scientific curiosity and institutional caution. While some researchers explore AI’s internal mechanics, the broader field remains skeptical about the potential for machine consciousness, framing it as a simulation rather than true awareness.
This content was automatically generated and/or translated by AI. It may contain inaccuracies. Please refer to the original sources for verification.