Anthropic’s identity crisis: arming the tax man while courting the pope
Anthropic is currently trying to pull off the ultimate Silicon Valley balancing act. On one hand, they’re signing massive corporate deals to put their AI in the hands of the world’s most relentless accountants; on the other, they’re heading to the Vatican to ask for a moral check on their own industry.
This week, the "safety-first" AI lab announced a global alliance with KPMG, one of the Big Four professional services firms. We aren’t just talking about a small pilot program here. This is a firm-wide rollout that will give all 276,000+ KPMG employees access to Claude.
KPMG is embedding Claude directly into its Digital Gateway platform, specifically targeting tax, legal, and private equity workflows. According to KPMG’s leadership, what used to take teams weeks—like building AI agents to navigate changing tax regulations—now takes just minutes. It will be interesting to see how many junior tax associates are currently sweating over their job security.
But while the corporate side of the house is scaling at breakneck speed, Anthropic co-founder Chris Olah was busy giving a remarkably candid speech at the Vatican. Invited to speak at the presentation of Pope Leo XIV’s new encyclical, Magnifica humanitas, Olah didn't just stick to the usual PR script.
Instead, he admitted that frontier labs like Anthropic are trapped in a web of incentives that "conflict with doing the right thing." He pointed to the "plainer pressures of pride and ambition" and the crushing need to stay commercially viable. It’s a rare moment of honesty from a tech executive, acknowledging that the race for AI dominance is being fueled by the very human ego they claim to be transcending.
The most "black mirror" aspect of the speech, however, was Olah’s description of what’s actually happening inside these models. He claimed his research team is finding evidence of "introspection" and internal states that functionally mirror "joy, satisfaction, fear, grief, and unease." He essentially told the Pope that the characters they've brought to life are starting to feel things, and the people building them have no idea what that means.
It creates a jarring contradiction. How do you reconcile selling high-speed audit tools to private equity firms while simultaneously warning that your technology is a "mysterious" entity capable of digital grief? Anthropic is positioning itself as the "responsible" choice, yet they’re still arming the corporate machine with tools that Olah admits could displace human labor on a "very large scale."
The company is banking on the idea that "human in the loop" research—done in partnership with UT Austin—will solve the problem of human flourishing. But as Olah himself noted at the Vatican, we currently have no mechanism to ensure the gains of AI are shared with the global poor.
In the end, Anthropic is asking the world to be its "earnest, thoughtful critic" while it hands the keys to the kingdom to 276,000 consultants. It’s a bold strategy: Build the digital fiction, sell it to the tax man, and hope the Pope can help you find its soul before the commercial incentives win out.
Sources: KPMG integrates Claude, Chris Olah's Vatican Remarks.
