There's a new news site on the block called The Wire, and it has a very specific editorial mission: tear down anyone who's skeptical about AI. But according to a new investigation from the watchdog group Model Republic, the site is almost entirely fabricated by AI — and may be a front for an OpenAI-backed political action committee.
Model Republic started digging after a reporter got forwarded a suspicious email The Wire had sent to a nonprofit called Encode. The sender didn't seem to exist.
The Wire launched on December 29, 2025, and in less than four months, it's published 94 long-form articles covering AI policy, Senate races, pharmacy reform, nuclear energy, and crypto regulation. There's zero information about who runs it or how it was set up.
Model Republic ran the articles through an AI detection tool called Pangram, which claims a near-zero false positive rate. The result: 69 percent of the 94 articles were flagged as fully AI-generated, and another 28 percent as partially AI-generated. Only three were classified as human-written.
The evidence doesn't stop at detection tools. In the site's source code, Model Republic found a form for creating new articles that prompts users to enter "background information for the AI to generate questions and write the article." Another section asked for "questions for the AI interviewer to ask." Someone appears to have been using this form to auto-generate stories.
The Wire's API also includes a feature that lets AI flag problems and suggest fixes. Of the 94 articles, 42 were marked "needs revision" by an automated review check — but they were all published anyway.
One article quotes a Harvard Business School professor. Others cite anonymous sources described as Republican staffers or industry insiders — but in some cases, those quotes appear to have been lifted from other news sites.
The Wire is virtually invisible on Google, and on X it's been linked only four times, all by relatively obscure accounts. Two of those links came from Patrick Hynes, the president of a Republican PR firm called Novus Public Affairs. Model Republic suspects Hynes is behind the site.
One clue: Hynes is based in New Hampshire, and The Wire has a unusually high number of New Hampshire-focused articles. Another: the site's coverage often overlaps with Novus's client work. For instance, 10 days before President Trump signed a major pharmacy benefit manager reform, The Wire published a scathing piece about PBM abuses. The reform was pushed by PhRMA, the pharmaceutical industry lobby, which spent heavily on lobbying — and PhRMA is a client of Novus. The article's source log even included a note labeled "PhRMA statement."
The Wire also goes after anti-AI voices. In one article, it attacked AI safety advocate John Sherman for suggesting on his podcast that "if people understood the risks of AI, they'd burn down data centers." Rather than just reporting the quote, The Wire listed every client of Sherman's video production and consulting firm and asked each one if they were aware of his remarks and planned to continue their relationship. The article claims all organizations declined to comment, but one "suggested it had privately severed ties."
About 15 percent of The Wire's articles cover AI, and many echo the talking points of companies opposed to AI regulation. Model Republic says "more than a third of The Wire's articles read like advocacy pieces for specific interest groups rather than journalism."
Many of those articles align with the messaging of Leading The Future, a super PAC (political action committee) primarily funded by OpenAI president Greg Brockman and investor Andreessen Horowitz, and created under the guidance of OpenAI's chief political operator Chris Lehane. OpenAI's political operation relies heavily on a Republican consulting firm called Targeted Victory — which happens to be a client of Novus Public Affairs. And all of this leads Model Republic to allege that "OpenAI's super PAC may be using The Wire to push a political agenda under the guise of independent journalism."
OpenAI has recently published a set of principles calling for democratic decision-making around AI — a stance that looks increasingly ironic given the alleged astroturfing campaign.