OpenAI and the Mystery of AI Journalists: An Unsettling Investigation

A News Site Under the Influence of AI

A news site, known as The Wire by Acutus, is at the center of a controversy for using artificial intelligence agents posing as real journalists. This practice, highlighted by an investigation from Model Republic, published by The Midas Project, reveals indirect links to OpenAI. The site, launched on December 29, 2025, appears to have no identifiable human contributors. According to Pangram, an AI detection tool, about 97% of the site's content is generated, in whole or in part, by artificial intelligence.

An Automated Editorial Process

The investigation uncovered, by exploring the site's source code, that human intervention is minimized. The site uses specific fields to provide the AI with basic information necessary for content generation, such as creating questions, writing articles, and suggesting questions for AI-conducted interviews. The site's RSS feed describes a largely automated editorial process, where out of five revision steps, only one requires human intervention. This process takes an average of 44 seconds, suggesting a more symbolic than real proofreading.

A field named "aiOriginalText" allows for comparison between the text generated by the AI and the proposed modifications, illustrating a new stage in the use of AIs for content creation. The site goes further by using AI agents to conduct real interviews with human experts. For example, Nathan Calvin, vice president and legal advisor of the organization Encode, received an email from a supposed journalist named Michael Chen, who does not exist online. The address used was a simple generic contact from the site, and the site's code mentions terms such as "journalist agent" or "AI interviewer."

Indirect Links to OpenAI

The investigation revealed that several articles from the site were shared by Patrick Hynes, an influential figure in public relations, close to the Republican Party. Behind him, connections to significant funding structures in the AI field, involving personalities linked to OpenAI, have been identified. Although there is no formal evidence of direct involvement, the investigation suggests that this type of platform could be used to disseminate pro-AI ideas under the guise of independent journalism.

The site does not only promote AI; it also targets its critics. One article attacks John Sherman, a researcher and advocate for AI safety, by contacting the organizations he collaborates with to question their relationships.

A New Era for AI Journalism?

Even without confirmed links to OpenAI, the use of automated "journalists" to conduct interviews and publish pro-AI content raises many questions. While the use of AI in newsrooms is already a reality, often to assist with proofreading or brainstorming, this situation seems to mark a more radical turning point. This evolution comes as OpenAI expands its presence in the media, notably through the recent acquisition of the highly followed TPBN show in Silicon Valley. This strategy echoes that of other tech giants, who have already invested in media or launched influential publications.