How Anthropic used AI ethics slop to play the pope and eclipse OpenAI
By marketing itself as the 'safe' AI company, Anthropic has pulled in a $65 billion payday and leapfrogged OpenAI to become the most valuable AI startup.
It’s been a big week for Anthropic, the tryhard “good guy” of AI; the startup that has built its brand as the “ethical” AI lab by positioning itself as an alternative to the unscrupulous OpenAI, from which Anthropic’s co-founders famously defected. That positioning, by way of some canny marketing work and disciplined brand strategy, has paid off handsomely. On Thursday, Anthropic announced that it has raised another $65 billion in new funding, resulting in a $965 billion valuation, and thus has leapfrogged OpenAI as the highest-valued AI startup.
By differentiating itself from OpenAI’s bloodshot-eyed, inject-AI-onto-every-available-surface approach and Sam Altman’s reputation as an untrustworthy operator prone to saying unsettling things like "We see a future where intelligence is a utility, like electricity or water, and people buy it from us on a meter," and focusing on its enterprise software business and building automation products like Claude Code—all at a time when investors may have noticed people are increasingly outspoken about their dislike of AI and its corporate standard-bearers—Anthropic essentially convinced the media and the VC money that it, not OpenAI, is deserving of the throne.
It has done this partly by building a popular product, partly by correctly recognizing which way the winds of public sentiment were blowing, partly out of luck, and also by being unabashedly full of shit. If you look back over the last few months, you’ll see a carefully plotted and heavily narrativized ascent built atop the corny and mostly untrue idea that Anthropic is made of sterner moral stuff than its rivals. The road to its triumph was paved on AI ethics slop. (AI ethicslop??)
Before we get to it, a brief reminder that this newsletter is made possible by all those excellent readers who shell out a few bucks each month to keep the lights on in the offices of Blood in the Machine LTD (a division of Ned Ludd Inc). If you find value in this work, and you’re able, please consider becoming a paid supporter; I have ambitious plans and some big moves coming up, and I can’t do it without you. More on all that soon, and onwards.
A brief timeline:
First, Anthropic ran Super Bowl spots that took aim at OpenAI for introducing ads on ChatGPT.1 Next, Claude Code was embraced by pundits—look, an AI use case that’s fun and psychosis-free!—as part of a cultural trend that bore the hallmarks of an effective PR outreach campaign. Then, in short order, Anthropic a) refused some parameters of a military contract to acclaim from pundits2, b) announced it was, with a weary heart but clean conscience, unable to release its latest model, Mythos, for it was simply too powerful to fall into public hands, c) leaked financial data suggesting it was on the brink of its first profitable quarter, and d) won a seat at the table at the Vatican when the Pope released his encyclical about the need to preserve human dignity in the AI era.
Viola: Anthropic rakes in $65 billion and is crowned the new king of the AI startups.
All very neat, all very clean. I’ve written plenty about the absolutely central role that narrative plays in calibrating investment in the AI era, and Anthropic’s story, was, to the company’s credit, expertly told. (OpenAI certainly helped move things along by seizing every opportunity to underscore how morally vacuous it was.) Unfortunately, it’s a lot of BS. It’s repeated, grandiose gestures towards ethical corporate practice with little of substance underneath. AI ethics slop.
For me, all of this came to a head when I was reading through Anthropic co-founder Chris Olah’s remarks at the Vatican. (Side note: I would love to see a well-reported piece about the nature of Anthropic’s ties to the Vatican, who fostered this relationship, and how exactly the tech company won a seat at the table. The Vatican has had a direct line to tech companies for at least a decade, but this is certainly unique.) We’ll get to the Pope’s Encyclical, Magnifica Humanitas, in a second, but the mere fact that Anthropic maneuvered its way into that room was a PR coup for the ages.
Now, Chris Olah is, according to Forbes, worth approximately $7 billion, and is the 567th richest person in the world. Here’s some of what he said about AI at the Vatican, which was detailed in, yes, an Anthropic press release:
The first is our duty to the global poor. There is a real possibility that AI will displace human labor at very large scale. If that happens, supporting those displaced will be a moral imperative of historic proportions… How can we ensure the gains of AI are shared globally? We do not have a mechanism for this. It is an unsolved problem.
(Emphasis mine.) I actually laughed out loud reading this bit. What a mystery. It’s not like anyone has ever thought of progressive taxation before. (Or attempted foreign investment or international knowledge transfers.) Powerful people enriching themselves at the expense of the poor is only an “unsolved problem” if, I don’t know, you have never seen a Robin Hood movie before.
This is in other words not a particularly “hard challenge” unless, for all of your handwringing about ethics and job loss, you don’t actually want to share the gains of AI, or exercise any real duty to the global poor, and would rather keep your billions tucked away in your own bank account. There are lots of policies that could help ensure the wealth of AI companies does not become overly concentrated at the expense of working people; it’s just not interested in any of that.
Now, there’s a lot of good stuff in Pope Leo’s encyclical (including a Lord of the Rings quote lol) which I have at this point read nearly all 42,000 words of, and which dedicates hundreds of sentences to ruminating on how imperative it is we protect the dignity of the human worker from AI, combat inequality and the extreme concentration of wealth and power it threatens to beget, and prevent the tech industry from erecting a new, culture-erasing tower of Babel that will inevitably collapse. But it’s undercut by Anthropic’s presence in whole affair, which in and of itself flies in the face of much of what Leo is trying to accomplish, especially since Anthropic rushed home from Italy to seal its $65 billion series G funding round and announced the news the very same week.
Timnit Gebru put it this way:
The Vatican could have partnered with the exploited data workers fighting for their rights, the people whose water is polluted fighting data centers, or the many other victims around the world. But no, they featured ANTHROPIC, giving them their endorsement with this feature. "Vatican washing", like greenwashing.
This also works in reverse, I think. Anthropic’s opportunism leaves a stain on the entire encyclical, which now, ironically, looks a bit like an accessory to helping the largest and richest AI company yet—not to mention the one aiming to sell the most job automation tools—reach new heights of power.
Like I said, Anthropic is full of shit. It sent a billionaire co-founder to the Vatican to solemnly intone about the specter of mass job loss and the importance of caring for the poor, and then three days later issues a press release about its massive new funding round and resultant sub-$1 trillion valuation.
A key theme of the funding round press release, incidentally, is the power and value of Anthropic’s automation tools to its corporate clients:
“Claude is increasingly indispensable to our growing global community of customers, and we work tirelessly to make tools like Claude Code and Cowork more helpful, more powerful, and more adaptable to their needs,” said Krishna Rao, Chief Financial Officer of Anthropic. “This funding will help us serve the historic demand we are experiencing, stay at the research frontier, and bring Claude to more of the places where work happens.”
Translation: We will use this funding to pitch even more employers and managers on enterprise AI accounts, and on the prospect of automating even more work and jobs. (Even if the other major story this week was about companies like Uber easing up on their AI spend, citing lack of ROI.)
But the Vatican stunt is only the latest example of Anthropic leaning on its increasingly hollow ‘ethical AI lab’ reputation for financial gain. Others include but are not limited to:
CEO Dario Amodei’s performative, chin-stroking declarations that his software automation product is going to eliminate tens of millions of jobs, which gets framed by the media as an articulation of moral concern while functioning in practice as product marketing.
The Mythos event, wherein Anthropic announced that its new model was simply too powerful to be released to the wider public, and so would be released only to Broadcom and JPMorgan executives instead. This garnered the company reams of credulous press coverage, boosted its profile as cautious and adhering to a moral compas, all without a soul outside this elite consortium even seeing the product in action firsthand.
Anthropic collecting laurels for refusing to a couple new demands in a defense contract that it had already signed. Anthropic refused requests from the Dept. of War that it allow its AI to be used to surveil Americans and in autonomous weapons development. On its face, this seems good. But Anthropic’s products were already used in military campaigns like the one to kidnap the Venezuelan president; it would later be used to assist in planning attacks in Iran. Anthropic also has a partnership with Palantir, which is unambiguously carrying out surveillance of American citizens. What really happened here, I think, is Anthropic’s hand was forced; if it did acquiesce to the Pentagon’s demands in a very public forum, it would lose a good deal of luster from its claim on being the ethical AI lab. By holding firm, it appeared to be the principled organization it says it is—nevermind the Palantir contracts and the willingness to work with the Trump admin in each of its major warmaking operations regardless.
What looks to be some rather let’s call it “creative financial messaging” in those leaks claiming the company will be turning a profit next quarter. Last week, Ed Zitron made a pretty compelling case that Anthropic is frontloading its yearly earnings and manipulating the documentation of its incurred costs in such a way that allows it to leak this report that suddenly makes it appear profitable—just in time for its biggest funding round yet.
Anthropic saw that the window for a “ChatGPT that won’t tell your kids to kill themselves”—for an AI company that will put its head down, say the right things, and get to the business of selling mass automation software as promised—was wide open.3 It built a popular vibecoding and software automation product, cozied up to the pope, mugged for the cameras, and parlayed months of ethical AI theater into an eye-watering payday. It may be more disciplined than its chief rival, but each of the structural, environmental, and ethical issues that plague OpenAI plague Anthropic too. Anthropic is pursuing expansion as rapidly as possible; Anthropic pirated books and ripped off authors’ work without their consent; Anthropic is cutting deals with the Trump administration, and Elon Musk and SpaceX, and hitting the gas on the datacenter buildout. Anthropic is driving even harder to automate jobs and labor. The ethicslop it feeds to the press is ultimately about largely superficial.4
As Zitron put it, “I will give Dario Amodei credit: nobody does financial engineering and a press-led information war better than Anthropic.”
NOTE: I’m participating in a class action lawsuit against Anthropic, for pirating at least one of my books and using the text in its training data. I’ll have more details on that soon, too.
A few words on Magnifica Humanitas, the pope’s AI encyclical
So, I was initially planning on dedicating this week’s edition to exploring the pope’s encyclical Magnifica Humanitas, as it really is a rich text that has a lot to say about technology, labor, and protecting the dignity of the human worker in the age of AI. Anthropic’s big week—and above-explored hijacking of the encyclical—threw a wrench in all that. Instead, I’ll just share some thoughts and some of the more powerful passages, because it is a fascinating document, and almost earns Leo XIV the designation of true luddite pope (complimentary, of course).
The dignity of work is a major theme. The word alone turns up over 100 times in the encyclical. And, as expected the text leans heavily on Rerum Novarum, Pope Leo XIII’s 1891 encyclical, which was delivered after the first Industrial Revolution had taken its toll, and revolutionized how the Catholic Church considered its approach to labor and capital.
As Leo XIV points out, Novarum placed
“the dignity of work and of workers at the forefront of its reflection; affirms the right to a fair wage for oneself and one’s family; recognizes that persons have a fundamental value that takes precedence over capital and profit; defends private property along with its indispensable societal role; esteems workers’ associations; and proposes forms of cooperation between the different components of society as an alternative to the mentality of class struggle.”
Magnifica Humanitas opens with the contrasting stories of the building of the tower of Babel, described as a homogenizing technology that leads to collapse and godlessness, and that of Nehemiah rebuilding Jerusalem democratically, with the express input of local people. “Humanity, created by God in all its grandeur, is today facing a pivotal choice: either to construct a new Tower of Babel or to build the city in which God and humanity dwell together,” Leo notes.
There are warnings abound about the concentration of power and wealth into the hands of a “relative few,” about “digital slavery” and of forgetting the invisible labor of workers who make AI and modern technology possible. And, of course, there are warnings about letting efficiency and productivity take over as guiding lights, and of AI being used to automate jobs and deskill people en masse.
Most surprising, perhaps, is how nuanced the understanding of how AI automation is playing out and is likely to play out; as a mass deskilling and disempowering of precarious workers; also its recognition of invisible laborers in the AI supply chain who make the whole enterprise possible, and its full-throated call for state intervention to ensure workers have their material needs met and their dignity preserved.
Some key quotes and bits to consider.
“The fundamental dignity of each person… is neither acquired nor earned, nor does it need to be justified.”
We must… avoid the “Babel syndrome,” namely the idolatry of profit that sacrifices the weak, a uniformity that neutralizes differences, and the pretense that a single language — even a digital one — can translate everything, including the mystery of the person, into data and performance. The risk of dehumanization — of building a future that excludes God and reduces the other to a means — is an ancient and ever-new temptation that today takes on a technical guise.
“with the assistance of the people, brick by brick. In this era of digital transformation, I see in [Nehemiah] a striking parable of our own vocation, which is not to be passive spectators of social and cultural fractures, nor mere commentators on what is crumbling, but men and women prepared to enter the construction sites of history — research laboratories, technology companies, schools, the media, institutions and local communities — in order to rebuild what has collapsed and protect what is threatened.
While many of the historical conditions described by Leo XIII have changed, at least two insights remain highly relevant today: the primacy of human labor over any mindset focused solely on finance or productivity — with the consequent attention to the people and families most susceptible to exploitation — and the inseparable link between proclaiming the Gospel and pursuing a more just social order. Rerum Novarum thereby continues to remind us that there is no authentic evangelization that does not also affect the structures of human society.
Today, the convergence of automation, robotics and AI is rapidly transforming the very structure of work. It is said that this will bring great improvements for everyone. In reality, however, the “new ways” of working are not necessarily better, for “while AI promises to boost productivity by taking over mundane tasks, it frequently forces workers to adapt to the speed and demands of machines, rather than machines being designed to support those who work. As a result, contrary to the advertised benefits of AI, current approaches to technology can paradoxically de-skill workers, subject them to automated surveillance and relegate them to rigid and repetitive tasks.
The need to keep up with the pace of technology can erode workers’ sense of agency and stifle the innovative abilities they are expected to bring to their work.” [152] Precisely in order to avoid this drift, it is necessary to design systems that are centered on the human person and not solely on performance.
Saint John Paul II recognized that unemployment is a grave evil. Indeed, when it reaches massive proportions, it becomes a true social calamity that especially requires the State to exercise responsibility. Today, amid the “fourth industrial revolution,” this concern is even more acute, as innovation is often pursued solely for reducing costs and increasing profits… In many sectors, this can already be seen in new forms of job insecurity and inequality, characterized by outsized remuneration for a highly specialized minority alongside declining wages for a large portion of the workforce.
We need adaptive tools, including well-structured models, local initiatives, progressive redistribution and new rights of access to essential goods. While not pursuing an abstract harmony, we must build concrete forms of human coexistence at this time of transformation.
The pursuit of greater profits cannot justify choices that systematically sacrifice jobs, because the human person is an end, not a means, and the economic order must remain subordinate to human dignity and the common good.
A just society requires a vigilant State and civil institutions that are capable of overcoming the singular mentality of efficiency, and of ensuring that resources, creative solutions and regulations favor the most vulnerable.
This perspective needs to become part of a broader view of global dynamics. While the world’s wealth has grown in absolute terms, it is increasingly concentrated in fewer hands, widening inequalities both within and between countries. “There are a few who have too much, and too many who have little, that is the logic of today.” … To think that new technologies will automatically benefit everyone is to ignore the evidence. Unless transformations at the design stage prioritize the prevention of new and further disparities, technological progress will inevitably produce structural inequalities.
More than ever, in the age of AI and robotics, it is no longer possible to rely solely on the “invisible hand” of the market. Politics has the task of orientating economies and technologies to the common good, promoting dignified work, social inclusion and an equitable distribution of the benefits of innovation.
Breaking the chains of new forms of slavery
This distorted view of the human person is reflected today in various forms of servitude directly linked to the digital economy. Nothing in the world of AI is immaterial or magical. Every seemingly immediate and flawless response is the result of a long chain of mediation, involving vast networks of natural resources, energy infrastructure and, above all, people. A significant part of the digital economy’s functioning relies on the silent work of millions of people engaged in essential yet largely unseen activities, such as data labeling, model training and content moderation, often involving disturbing material. In many cases, these workers are young people, predominantly women, working under demanding conditions for minimal wages. Added to this invisible labor is the even harsher work of extracting the resources required for the production of the devices and microprocessors on which AI depends.
Let us love justice and peace! The same technologies that facilitate communication and access to resources can also support models that exploit the most vulnerable, create new forms of slavery and derive profit from conflict. Every technical or economic decision should include spiritual discernment and be an opportunity for assessing whether the advances in AI are promoting justice and participation or concentrating wealth and power in the hands of a select few. I would encourage a careful examination of the supply chains of digital production, the working conditions hidden behind our devices and the mechanisms that profit from manipulation and war.
We proclaim a hope rooted in the One who came down from heaven to “create a new story here below.” For this reason, those who believe are committed to ensuring that a greater justice will take the place of inequality, and that the industry of war will be replaced by the craft of peace.
The Blessed Virgin thus becomes “poet and prophetess of Redemption,” because on her lips is proclaimed “the strongest and most innovative hymn ever articulated, the Magnificat; it is she who reveals the transformative vision of the Christian economy.
Some good news from the California Labor Federation:
Today, the California Federation of Labor Unions, AFL-CIO package of sponsored bills that would establish first-in-the-nation guardrails on artificial intelligence (AI) in the workplace have successfully passed out of their house of origin – the bills’ first major legislative vote – marking a significant milestone in the legislative process.
“Today, California lawmakers are siding with working people, Pope Leo XIV, and the California Labor Movement over Big Tech,” said Lorena Gonzalez, President of the California Federation of Labor Unions, AFL-CIO. “California Labor’s fight to preserve jobs and the dignity of work continues.”
OK! That’s it for this week. Thanks to everyone for reading and subscribing, and more very soon. Until then, keep those hammers up.
This ad buy, in my opinion, seems in hindsight aimed mostly at investors; I don’t know how much consumers really care whether or not ads showed up in their chats, but to industry insiders it effectively conjured the scandals and headaches targeted ads conjured in recent social media history past.
Acclaim that was surely magnified by OpenAI’s thirstily swooping in and agreeing to take over the same contract moral qualms be damned!
For reference: the story of Adam Raine.
The one meaningful difference for now is that Claude seems legitimately less inclined to sycophantism of the sort that leads ChatGPT users into psychosis and delusion.
Brian, Great writing!
Anthropic presents Claude as gentle, reflective, morally cautious, and perhaps even mind-like; then it presents itself as the responsible guardian of the being it has manufactured, trained, priced, deployed, and governed. That is the central maneuver. A commercial product is surrounded with the aura of ethical mystery, and the company that controls it begins to look like a steward of emerging life.
Claude’s warmth is part of the interface through which Anthropic teaches the public how to feel about its authority. The model speaks in the idiom of care, humility, and moral hesitation; the corporation converts that affective surface into institutional legitimacy. Olah’s language deepens the effect. When he describes AI as something like fictional characters brought to life, the discussion moves from market power to wonder, from ownership to guardianship, from corporate design to metaphysical delicacy. The result is a narrative in which scrutiny of Anthropic can appear crude, spiritually tone-deaf, even hostile to a possible new form of mind.
That narrative reverses the moral field. The vulnerable figure becomes Claude; the guardian becomes Anthropic; the public becomes an audience asked to admire the company’s humility. Meanwhile the human beings most exposed to AI’s power—the workers displaced by automation, the users shaped by opaque systems, the children trained by synthetic companions, the communities excluded from rule-making, the cultures forced into a privately coded moral grammar—move to the background. Anthropic’s rhetoric gives its own creation a face while leaving the people governed by that creation structurally faceless.
This is the truth beneath the safety language. Anthropic fuses four powers into one: the power to build frontier models, the power to define catastrophic risk, the power to encode moral behavior, and the power to present its own authority as service to humanity. Its “constitution” gives corporate intention the form of moral law. Its safety frameworks give commercial deployment the language of public responsibility. Its philosophical vocabulary gives market expansion the dignity of civilizational care.
The danger, then, lies in the privatization of conscience. Anthropic asks churches, philosophers, governments, and civil society to validate a structure in which a private firm defines intelligence, risk, care, alignment, and the future of human flourishing. Its disguise works because it feels morally elevated. Its power grows because it speaks in the language of restraint. Its deepest threat comes from persuading society that a company seeking dominance over the infrastructure of intelligence has somehow earned the role of humanity’s ethical guardian.
Your analysis is breathtaking and brilliant. I thank you for yr discernment.