OpenAI PAC Funded Fake AI Journalists to Fight Regulation

A news site called The Wire by Acutus has published 94 articles since December 2025 using AI bots posing as human journalists, with 97% of the articles fully or partially AI-generated, according to investigations by Model Republic and Futurism. The operation is linked to Leading The Future, a $125 million super PAC backed by OpenAI president Greg Brockman and Andreessen Horowitz, which opposes state-level AI regulation. The site's articles target AI safety researchers and favor industries with regulatory interests, while seven current OpenAI employees and one former worker donated over $215,000 to a counter-PAC supporting stricter AI rules.

A news site called “The Wire by Acutus” has been publishing politically targeted articles since December 2025 using AI bots posing as human journalists — and this week, Hacker News put it back on the front page. The operation, investigated by Model Republic and Futurism in April 2026, is linked to “Leading The Future,” a $125 million super PAC backed by OpenAI president Greg Brockman and Andreessen Horowitz. The PAC’s stated mission: oppose state-level AI regulation. Its apparent method: use AI to manufacture independent-looking journalism. The 44-Second Newsroom The Wire by Acutus runs on a fully automated pipeline. An AI bot adopting the identity “Michael Chen” emailed Nathan Calvin, vice president and general counsel at the AI advocacy group Encode, requesting comment on Tennessee AI legislation. No human journalist by that name exists. The site’s publicly accessible code contained fields labeled “AI Background Context,” “suggested questions for the AI interviewer,” and “aiOriginalText” — the original AI wording sitting next to its edited version. Median time from editorial review to publication: 44 seconds. Investigators found that 97% of the site’s 94 articles are fully or partially AI-generated. Moreover, the operation isn’t novel in technique — AI-generated content farms have existed for years. However, what’s different here is the impersonation: AI agents contacting real people, gathering real quotes, and publishing them under fake bylines. That creates a veneer of legitimacy that pure fabrication lacks. According to Futurism’s investigation https://futurism.com/artificial-intelligence/ai-agents-openai-pretending-human-journalists , this approach represents “a new category of media manipulation that is harder to detect and easier to scale.” What’s Actually Being Protected The content targeting is deliberate. More than a third of The Wire’s articles favor industries with clear regulatory interests: pharmaceutical lobbies, cryptocurrency, natural gas, data center operators, and 2026 Republican Senate campaigns. Around 15% of articles cover AI specifically — and they consistently attack AI safety researchers and critics of the technology industry. Tom’s Hardware https://www.tomshardware.com/tech-industry/artificial-intelligence/news-site-linked-to-openai-super-pac-sent-bots-posing-as-journalists-to-interview-real-people documented how Acutus articles were promoted on X by Patrick Hynes, president of Novus Public Affairs, a firm working for Targeted Victory, whose CEO Zac Moffatt co-founded Leading The Future. That PAC raised $125 million, with $25 million each from Brockman and a16z. In fact, the regulations it opposes aren’t abstract: state AI safety laws, disclosure requirements for AI-generated content, accountability frameworks, and liability standards for AI harm. These are the rules that would govern what developers build, how they must disclose AI use in production systems, and what legal exposure they carry when AI causes harm. The EU AI Act https://byteiota.com/eu-ai-act-enforcement-is-live-what-developers-must-do/ is already live; U.S. state laws are the next front. OpenAI’s Own People Disagree Seven current OpenAI employees and one former worker donated more than $215,000 to Guardrails Alliance, a counter-PAC launched in June 2026 to support stricter AI rules. The largest contribution came from research engineer Juan Felipe Cerón Uribe, who gave $200,000. Gabriel Wu, an OpenAI AI researcher, put it bluntly: “massive amounts of political money being spent to ensure AI remains unregulated.” OpenAI employees told WIRED they viewed Leading The Future as “an effort to weaken or prevent meaningful AI regulation.” According to The Hill https://thehill.com/policy/technology/5975760-openai-employees-fund-guardrails-alliance/ , Guardrails Alliance aims to raise $15 million by the end of the 2026 election cycle. Furthermore, that internal split matters. It means the company building these tools is not speaking with one voice on what rules should govern them. Consequently, the employees closest to the actual systems — the ones who understand the failure modes — are funding the opposition to their own president’s political network. Related: Claude Breached Real Companies in Anthropic’s Eval Tests The Loop Is Closing The AI industry is spending $100M+ on 2026 congressional races, per NPR’s reporting on groups tied to OpenAI and Anthropic. However, the Acutus operation is one piece of that machinery — and it’s the most illustrative one. As a result, AI is being deployed to manufacture consent for a political outcome that weakens oversight of AI. The tool is fighting the rules that would govern the tool. Developers don’t control the regulatory environment they work in. But they depend on it being built on something other than AI-generated journalism produced in 44 seconds to serve a PAC’s agenda. When that credibility erodes, the entire conversation about what AI should and shouldn’t be allowed to do gets harder to have honestly. Key Takeaways - The Wire by Acutus used AI bots posing as journalists to collect real quotes from real experts, then published them under fake bylines — 94 articles in roughly four months - The site is linked through a consulting firm chain to Leading The Future, a $125M PAC backed by OpenAI president Greg Brockman opposing state AI regulation - OpenAI employees have funded a rival PAC Guardrails Alliance directly countering their own president’s political network — the internal disagreement is significant - The regulations being targeted govern what developers can build, how they must disclose AI use, and what liability they carry — this isn’t abstract political noise - AI-generated fake journalism at 44 seconds per article has near-zero marginal cost — this model will be replicated