Chris Olah warned that artificial intelligence development should not be controlled exclusively by technology companies, calling for stronger oversight from governments, religious institutions, and civil society groups.
Olah made the remarks during a Vatican event presenting the first AI-focused encyclical from Pope Leo XIV, where he appeared alongside the pope to discuss the societal impact of advanced AI systems.
The Anthropic co-founder said there was a “real possibility” that AI could displace human labor at massive scale and argued that supporting displaced workers would become “a moral imperative of historic proportions.”
AI Labs Face Conflicting Incentives
Olah said frontier AI companies operate under strong commercial and geopolitical pressures that may conflict with broader societal interests.
“Every frontier AI lab operates inside a set of incentives and constraints that can sometimes conflict with doing the right thing,” he said.
He argued that even well-intentioned researchers remain shaped by competitive dynamics, making external scrutiny increasingly important as AI systems become more powerful.
According to Olah, ethical questions surrounding AI extend far beyond engineering challenges and require participation from critics, policymakers, philosophers, and public institutions.
Vatican Expands Focus on AI Ethics
The event marked another step in the Vatican’s growing involvement in debates around artificial intelligence governance and ethics.
The Catholic Church has increasingly positioned itself as a moral voice on emerging technologies, particularly around automation, human dignity, surveillance, and economic disruption.
Olah welcomed the Church’s involvement, saying the questions raised by AI were “bigger than the AI research community.” He specifically highlighted three urgent concerns: large-scale job displacement, unequal global access to AI benefits, and the growing difficulty of interpreting advanced AI system behavior.
Anthropic Continues Emphasis on AI Safety
Anthropic has become one of the most vocal frontier AI companies around AI safety, governance, and deployment restrictions.
The company previously clashed with the administration of Donald Trump over military-related AI guardrails, particularly restrictions tied to autonomous weapons targeting and domestic surveillance applications.
Anthropic’s broader positioning has focused heavily on responsible AI deployment, interpretability research, and controlled access to high-risk systems such as its cybersecurity-focused Claude Mythos Preview model.
The company has also recently expanded enterprise adoption efforts through partnerships with firms including PwC and KPMG while continuing to publicly advocate for stronger governance frameworks around advanced AI systems.
Global AI Governance Debate Intensifies
Olah’s remarks arrive as governments, regulators, and AI companies increasingly debate how much oversight should exist around frontier AI development.
Companies including OpenAI have publicly supported the creation of international AI governance structures modeled after nuclear oversight organizations such as the International Atomic Energy Agency.
At the same time, concerns continue growing around labor disruption, concentration of AI power among a small number of companies and countries, and the difficulty of understanding increasingly autonomous AI systems.
Olah warned that AI development remains heavily concentrated in a small number of wealthy nations and questioned how the benefits of the technology could be distributed more globally.
