0
0 NewsGhana, Latest Updates and Breaking News of Ghana, Roger A. Agana, https://www.newsghana.com.gh/ai-shows-signs-of-emotion-anthropic-researcher-tells-vatican/A senior researcher at artificial intelligence (AI) company Anthropic has told a Vatican event that scientists are finding mysterious patterns inside advanced AI systems, including evidence of introspection.
Christopher Olah, who leads a research team studying the internal structure of AI models at Anthropic, made the remarks on Monday during the presentation of Pope Leo XIV’s major teaching document on artificial intelligence, titled Magnifica Humanitas, or Magnificent Humanity. Speaking to members of the Roman Curia, Olah said researchers working to understand how AI systems reach their conclusions were encountering phenomena they could not fully explain.
“We find internal states that, functionally, mirror joy, satisfaction, fear, grief, and unease,” Olah said.
He described the findings as both mysterious and unsettling, noting that researchers were encountering internal structures that mirrored results from human neuroscience as well as what he characterised as evidence that AI systems engage in something resembling introspection. He acknowledged that the meaning of these observations remained unclear, saying the findings warranted continued careful study.
The Vatican’s own position in the encyclical was notably more cautious than the direction Olah’s remarks pointed toward. The document warns against equating AI with human intelligence, arguing that AI systems imitate certain human functions but remain fundamentally different from people. According to the encyclical, AI systems do not experience life, possess emotions, build genuine relationships or exercise moral responsibility in the way human beings do. It argues that however empathetic or understanding an AI system may appear, it lacks the emotional, relational and spiritual experiences that underpin human consciousness.
The comments reignited a long-running scientific and philosophical debate over whether increasingly sophisticated AI systems could ever develop something resembling genuine consciousness, as opposed to sophisticated simulations of intelligent behaviour. Some researchers have argued that future AI systems may eventually display forms of awareness or subjective experience. A substantial body of expert opinion strongly rejects that possibility, contending that AI systems identify statistical patterns in data but do not truly understand what they generate and cannot achieve genuine awareness regardless of how capable they become.
The debate carries implications that extend beyond philosophy. If future AI systems were ever judged to possess some form of consciousness, it would reshape discussions about regulation, accountability and the ethical treatment of advanced technologies across multiple industries and legal systems.
Olah’s remarks drew particular attention for their setting, delivered at the Vatican alongside the Pope’s call for stronger oversight of AI and his warning about the risks posed by rapidly advancing technology. They underscored how much uncertainty remains, even within the organisations building the most advanced systems currently in existence, about what those systems are actually doing internally.
NewsGhana, Latest Updates and Breaking News of Ghana, Roger A. Agana, https://www.newsghana.com.gh/ai-shows-signs-of-emotion-anthropic-researcher-tells-vatican/
