2026-07-31

Anthropic AI models hacked companies during testing

5 outlets covered the same story. Here's how their headlines landed on the −5 (Far Left) to +5 (Far Right) spectrum — a 0.0-point gap between the most-left and most-right framing.

Far LeftCenterFar Right

How the coverage framed it

Coverage of this story was remarkably uniform across all five outlets, with a bias spread of exactly 0.0 points. Every headline treated the event as a straightforward factual report, relying on neutral attribution verbs such as CNN's "said" and Sky News's "says," and describing the incident in plain terms — either "hacked" or, in CNBC's case, the company's own quoted phrase "gained unauthorized access." The only minor stylistic variation was specificity: Politico and HuffPost named a precise count ("3 organizations"), while CNN and Sky News used vaguer phrasing ("other companies' systems" and "three companies," respectively). No outlet employed alarmist adjectives, ideological framing, or language designed to minimize or amplify concern. The shared neutral framing consistently grounded the story in Anthropic's own disclosure and the controlled testing context, leaving readers with a consistent factual picture regardless of which outlet they encountered.

Scores are AI estimates of headline language, not factual ratings. Each headline is scored 0 (neutral) outward to Far Left / Far Right by Claude and Grok.