AI Newsrooms Beat Human Reporters to the Punch
OpenAI
Anthropic
RuntimeWire, an AI newsroom operated by serial entrepreneur Ryan Merket, beat WIRED to a story from Black Hat by more than three hours. Using AI agents that draft, edit, and fact-check, the site publishes about 100 stories a day at a cost of $100 daily. Other similar projects exist, but experts question the sustainability and quality of AI-generated journalism.
An AI-operated news site called RuntimeWire scooped human reporters by publishing a story about OpenAI's agents at the Black Hat conference more than three hours before WIRED's coverage, despite having no writers on the ground. Its operator Ryan Merket captured the OpenAI executive's live post on X, fed the transcript to his AI agents, and published the article in about six minutes. RuntimeWire, which has been running since May, publishes roughly 80-100 stories daily by crawling court databases, forums, and social feeds, with minimal oversight and a budget of $100 a day. Merket sometimes lets an AI editor publish stories without his review if legal risk is low. Similar projects exist, like the agentic newsroom The Dissent operated by BlackRock analyst Dakota Carrasco. Experts like Nicholas Diakopoulos see this as an experimental phase, noting that AI chatbots often surface AI-written sources. Pete Pachal doubts AI can replace human sourcing but sees it as a natural evolution for certain types of news. Merket claims to follow journalistic ethics but has retracted stories as a favor to founders.
Source: Wired AI —
original