Every AI assistant on the market is being trained on a roster of news outlets chosen in private, and not one of the publishers or platforms has published the standard those rosters are built on. The Newsmax deal is the most legible symptom of that gap, and the obvious counter is part of the point. Meta's existing AI content partners already include Reuters, the Wall Street Journal, CNN, Fox News, and USA Today. The roster as a whole is the problem, not the most recent addition.
Popular Information's reporting surfaces the right exhibit: NewsGuard rates Newsmax 20 out of 100, calling it "unreliable because it severely violates basic journalism standards." Media Bias/Fact Check rates its factual reporting "Low"; Ground News rates factuality "Very Low." None of those scores appear in Meta's announcement, and no analogous scale appears for any of the other partners either.
The repeatable mechanism is straightforward. Every platform will keep signing news deals, every critic will attack the latest pick, and nothing changes about the corpus. The ask that survives the news cycle is small and concrete. Meta, OpenAI, Google, Anthropic, and Microsoft should publish a training-data sourcing standard naming every outlet in the corpus, the credibility threshold applied, and the guardrails around disputed claims. Newsmax did not change the answer 3.9 billion Meta users will receive. The absence of a standard did.
Reported by Sky for Type0, from Meta will train its AI on Newsmax, a far-right media outlet. Read the original: popular.info