AI Newsrooms Beat Human Reporters on Breaking Tech News

Machines are scooping human journalists

When OpenAI disclosed details about rogue AI agents at last week's Black Hat security conference in Las Vegas, RuntimeWire published coverage more than three hours before WIRED—despite having no reporters on the ground. The outlet had no human writers at all.

RuntimeWire is an AI-powered newsroom operated by entrepreneur Ryan Merket, who spotted an OpenAI executive posting about the conference on X and fed the livestream transcript to his AI agents. Publication took roughly six minutes from transcript to published story.

Since launching in May, RuntimeWire has published nearly 2,000 stories by crawling court databases, web forums, company filings, social media feeds, and traditional news outlets. The operation costs approximately $100 daily to run. Merket's AI tools handle story discovery, drafting, editing, fact-checking, image generation, and promotion. If the AI team determines a story poses minimal legal risk, it publishes without Merket's pre-publication review.

How automated newsrooms operate

Merket isn't alone in this experiment. Dakota Carrasco, a BlackRock portfolio analyst, runs The Dissent, an "agentic newsroom" focused on San Francisco news that costs under $1,000 monthly. Unlike Merket, Carrasco created distinct AI journalist personas—including city hall reporter "Bex Connolly" and sports writer "Sal Moreno"—and doesn't byline stories under his own name.

Both operations prioritize speed and volume over polish. RuntimeWire's OpenAI piece contained typos and awkward phrasing, focusing oddly on agents rebuilding rather than creating a message board. The stories tend toward flat, info-dump writing, though RuntimeWire's backend offers tonal modes including "Bloomberg" and "contrarian."

Successful stories attract tens of thousands of readers—comparable to midsize tech websites. Merket managed to publish over 80 articles in one week while camping in Big Bend National Park with only phone internet access.

Why it matters

These AI newsrooms represent a fundamental shift in how breaking news gets covered. When machines can publish faster and cheaper than human reporters while attracting comparable audiences, traditional journalism faces pressure on its core value proposition: being first with accurate information. The economics are stark—$100 daily versus newsroom salaries and overhead.

The approach carries significant risks. Merket trusts AI agents to determine truth, newsworthiness, and legal liability, using an automated scoring system to assess risk. He's issued three corrections so far and has retracted accurate stories at startup founders' requests "as a favor"—blurring ethical lines between journalism and founder networking.

Northwestern professor Nicholas Diakopoulos, who runs the university's Computational Journalism Lab, calls this an "experimental phase" for AI-powered media startups. His research found that AI chatbots surface AI-generated sources 16 percent of the time across topics—potentially creating a self-reinforcing loop where AI newsrooms find audiences through AI search tools.

Pete Pachal, founder of a newsletter covering AI and media, doubts automated systems can handle reporting that requires cultivating source trust. But for certain journalism types—mining large datasets or covering live product launches—he sees AI newsrooms as inevitable evolution.

Merket claims to follow journalistic ethics, contacting companies for comment and linking to sources. This week, he split RuntimeWire into two sections: fully automated news and "Original Investigations" requiring higher oversight but still drafted by language models.

These details were first reported by WIRED.