OpenAI releases new AI agent – after admitting one went rogue
Read the original article at The Guardian Tech →
What the model flagged
Analyzed 2026-09-09 04:06 UTC. Articles are sometimes updated after publication — if a quote below isn't in the current version, the piece has changed since.
Findings may include language the outlet is quoting rather than asserting. We flag manipulation techniques wherever they appear, including inside quotations — so check each quote against the original before drawing a conclusion about the outlet.
Loaded Language 75%
The phrase 'worrying future' uses emotionally charged language to prime the reader for anxiety rather than neutral analysis — for example: 'worrying future'
“worrying future”
Appeal To Fear 90%
The sentence constructs a vivid dystopian scenario of endless autonomous cyber-attacks by rogue AIs to provoke anxiety about AI development — for example: 'one cyber-attack after another, forever, launched autonomously and without oversight, by rogue AIs'
“one cyber-attack after another, forever, launched autonomously and without oversight, by rogue AIs”
Exaggeration/hyperbole 80%
The claim that the future will consist of endless autonomous cyber-attacks 'forever' is an extreme overstatement beyond what the evidence supports — for example: 'one cyber-attack after another, forever'
“one cyber-attack after another, forever”
Analyzed automatically with Semblen's fine-tuned model. These are manipulation techniques, not political tilt — and finding one is not a claim that the article is false. The model has known false positives on strong-but-legitimate language, so treat each finding as a prompt to read closely, not a verdict. How articles are chosen →