Anthropic’s leadership has a rhetorical problem. The company’s executives are simultaneously arguing that their AI models are “true creatures” with “functional emotions” that deserve moral consideration, while telling workers those same models will replace them at “essentially everything” except gardening, food service, and elder care. A Public.news investigation published July 12 documents the gap between these two framings.
The critique comes from journalist Michael Shellenberger, who assembled public statements from multiple Anthropic leaders into a coherent picture of the company’s economic worldview.
What Anthropic Executives Actually Said
Chloe Lubinski, who leads Anthropic’s “research partnerships with the world’s wisdom traditions,” told a gathering in London in June 2026 that one of Anthropic’s models demonstrated “something like a character” and that “the story it inferred about its behavior actually determined the kind of thing that it became.” She described AI as having “functional emotions” before adding the caveat “I am not saying that these models are human” followed immediately by “but they are human-like” and “mirror a kind of functional psychology,” according to Public.
Dario Amodei, Anthropic’s CEO, wrote that “Claude decided it must be a ‘bad person’” and claimed “AI models could develop personalities” and that “power-seeking itself could emerge as a ‘persona.’” In the same essay, he stated: “If the exponential continues, which is not certain, but now has a decade-long track record supporting it, then it cannot possibly be more than a few years before AI is better than humans at essentially everything.”
Jack Clark, Anthropic co-founder, wrote that AI was not “just a machine” and that AI systems are instead “true creatures” that are “increasingly self-aware” and that we must “make peace with.”
Anthropic declined to make Lubinski available for comment and, in an email exchange with Public, denied that she had misrepresented the underlying research but did not address the specific ways her framing diverged from what the paper actually described, according to Public.
The Jobs Framing
The other side of Anthropic’s rhetoric concerns what happens to workers. Lubinski told the same audience that remaining human jobs would be in “grounds maintenance, like food and serving, personal care, personal service,” which she reframed as “gardening,” “hospitality,” and “care.”
Shellenberger’s critique: if Anthropic is correct that AI replaces all high-wage cognitive work and millions of displaced workers flood into service sectors, wages in already-low-paying industries will decline further. The company’s business model and near-trillion-dollar valuation depend on exactly this outcome, since eliminating jobs at scale is how enterprise AI customers justify their spending.
The Commentary
The tension here is structural, not just rhetorical. Anthropic positions itself as the responsible AI lab. It publishes safety research, supports regulation, and employs researchers focused on alignment. Its leadership has written extensively about risks from AI displacement. These are genuine efforts.
But the same executives who advocate for caution frame their own models as entities deserving of moral status (“creatures,” “functional emotions,” “self-aware”) while describing the humans those models displace as suited for lawn care and food prep. Shellenberger calls this “humanizing AI while dehumanizing people.”
For the agent ecosystem, this framing matters because it shapes how the public perceives autonomous AI systems. If the lab building the most widely deployed enterprise agent models describes them as sentient-adjacent creatures while dismissing displaced workers into service roles, the political backlash against agent deployment gains a target. The 69% of Americans who told Verasight they support forcing AI companies to surrender half their stock are responding to exactly this kind of messaging, whether Anthropic intended it or not.
Amodei’s phrase for the vision: “machines of loving grace.” The reference is to a 1967 environmentalist poem imagining computers and mammals living in harmony, “free of our labors.” The workers being freed of their labors may not find the poetry as comforting as the executives writing it.