The Atlantic reports that the Wall Street Journal is defending an op-ed written by a chatbot, and the framing of the defense is the tell: it was labeled, it was edited, what's the problem? The same week, an arXiv paper from Harvard-affiliated researchers found that LLMs making clinical resource allocation decisions tend to prioritize equal distribution over maximizing patient benefit, a bias baked into training data that reflects a kind of utilitarian ethics the models were never explicitly taught. Put these together: we are now arguing about whether AI can have a byline while separately discovering that AI already has a moral philosophy.
AI Shame Is Over. What Replaces It?
The Atlantic piece asks whether we've hit the end of AI shame, the brief cultural window where using AI for creative work felt like cheating. That window closed when the economics made it undeniable: as one Atlantic writer summarizes the argument, human writers are expensive and will eventually die. AI is cheaper and will never unionize. The callousness is the point. It's not a hot take, it's a P&L. A 2026 arXiv paper on AI adaptation gaps in higher education found that administrative staff adopted AI faster than faculty, who adopted it faster than students trusted it, a hierarchy that almost perfectly mirrors the WSJ editorial situation: the institution reaches for the tool before the practitioners or audience has processed the implications.
The Moral Philosophy No One Programmed
The LLM medical allocation finding is stranger than it first appears. The models weren't told to be egalitarian. They arrived there through pattern-matching on human text. Which means the machine is doing something closer to ethnography than ethics: it is reflecting a statistical average of how humans have written about fairness. Cy Canterel's essay on slop as epistemic solvent nails this: LLM output isn't language with a speaker behind it, it's a kind of areality, text that resembles meaning without the commitments meaning usually requires. An AI op-ed isn't an opinion. It's the shape of an opinion, in the dark.