Chris Olah’s Vatican speech placed one of Silicon Valley’s most prominent AI safety researchers beside Anthropic cofounder and interpretability research lead on a stage at the Holy See, where he told the assembled audience that computer scientists are the wrong people to govern artificial intelligence.
Olah, a self-declared atheist and a billionaire cofounder of Anthropic, sat alongside Pope Leo XIV on 25 May 2026 as the pontiff released his encyclical, Magnifica Humanitas, calling for a measured and vigilant approach to AI development.
‘Some might believe that matters of AI are best handled by computer scientists like myself,’ Olah said in his prepared remarks. ‘They are mistaken.’
What Olah said in his Chris Olah Vatican speech
Olah acknowledged the strangeness of his presence at the Vatican, opening his remarks by telling the audience that what he was about to say ‘may sound strange coming from the co-founder of an AI company.’
His core argument: no matter how sincerely those inside the industry intend to do the right thing, they will always be shaped by commercial incentives. Outside voices, including the Catholic Church, scholars and governments, must therefore keep the industry’s moral obligations in focus.
‘No matter how sincerely any of us intend to do the right thing, and I believe many of us do, we will always be influenced by those incentives,’ he said.
The position puts Olah at odds with figures such as venture capitalist Marc Andreessen, who argued in his 2023 Techno-Optimist Manifesto that ‘trust and safety’ and ‘tech ethics’ formed part of a demoralization campaign against technology. It aligns, however, with Anthropic’s stated mission, which emphasises safety and does not shy away from publishing research on AI risks.
From Toronto dropout to Anthropic cofounder
Olah was raised in Toronto, Canada, as a devout evangelical Christian. He became an atheist at 15. He enrolled at the University of Toronto to study mathematics but dropped out about a year in.
In 2012, he received $100,000 through the Thiel Fellowship, a programme created by PayPal cofounder Peter Thiel for talented young people who forgo a traditional four-year degree.
He joined Google Brain in 2015 as an intern, eventually becoming a research scientist. During those three years, he helped build tools to visualise the internal workings of neural networks, contributing to a field that would become known as mechanistic interpretability. The work was not widely prioritised at the time, when most researchers were focused on making AI more powerful.
Among his contributions was a paper titled The Building Blocks of Interpretability, which offered an early window into how neural networks derive complex concepts from simpler building blocks.
From 2018 to 2020, Olah led OpenAI’s interpretability team, where he worked on the Circuits project and co-led the discovery of multimodal neurons in CLIP, OpenAI’s model for connecting text and images. His team found that certain neurons fired in response to the same concept whether it appeared as a photograph, a drawing or as text, suggesting artificial neural networks may operate similarly to the human brain.
In 2020, Olah was among seven OpenAI employees, including chief executive Dario Amodei, who left the company over AI safety concerns. The group later cofounded Anthropic, which was valued at $965 billion after a recent funding round and this week confidentially filed for an initial public offering.
Olah’s net worth now stands at just under $8 billion, according to the Bloomberg Billionaires Index.
In 2024, Time named him to its TIME100 AI list. ‘If we could really understand these systems,’ he told the magazine, ‘we might be able to go and say when these models are actually safe. Or whether they just appear safe.’
Pope Leo XIV’s encyclical, released on 25 May 2026, calls for consideration of humans over machines and a vigilant approach to the technology’s development.
