Christopher Olah, co-founder and head of interpretability research at Anthropic, delivered a speech at the Vatican's Synod Hall during the presentation of Pope Leo XIV's first encyclical, "Magnifica humanitas." In this prestigious setting, he emphasized for the first time that the development of advanced artificial intelligence (AI) cannot be left solely to the laboratories creating it. This is reported by Ixbt.com. reports .
Christopher Olah noted that even the most conscientious researchers work under the pressure of commercial, political, and personal interests that may conflict with the public good. Therefore, he believes that external oversight by religious leaders, states, and civil institutions is necessary. The second part of the speech addressed risks to the labor market: according to Anthropic's estimates, AI could displace people from jobs on such a scale that the economy might not withstand it without additional support.
Anthropic's involvement at the Vatican was one of the most significant steps of 2026. The company opened its office in Milan and is now participating in shaping the most significant religious document since the 1891 encyclical "Rerum novarum." The team led by Olah works on understanding the internal mechanisms of advanced AI models, which is crucial for the company's reputation regarding safety issues.
At the same time, Anthropic is facing pressure from US officials: in the spring, the company was excluded from classified military projects due to its internal restrictions, and the expansion of its Mythos autonomous system, capable of finding vulnerabilities in bank cybersecurity, was blocked. Olah's speech was a direct response to these issues, reaffirming that companies cannot self-regulate and that a system of external oversight is required.
Read “Zamin” on Telegram!
