How an AIPAC endorsement became a liability
Read the original article at Vox →
What the model flagged
Analyzed 2026-08-13 04:01 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 65%
The word 'shady' is an emotionally charged term implying dishonesty or corruption without providing specific factual basis — for example: 'quite shady'
“quite shady”
Card Stacking 70%
The sentence presents AIPAC's framing as a straightforward fact without acknowledging the contested nature of equating criticism of Israel with antisemitism, omitting the counterargument that such criticism can be legitimate — for example: 'tying criticism of Israel to antisemitic sentiment'
“tying criticism of Israel to antisemitic sentiment”
Conspiracy Appeal 89%
Hidden plots, secret coordination, or shadowy actors invoked to explain away evidence.
Exaggeration/hyperbole 75%
Claiming the US is funding 'the destruction of an entire region' is an extreme overstatement that goes beyond what evidence typically supports — for example: destruction of an entire region
“destruction of an entire region”
Name Calling 65%
The label 'Democratic socialists' functions as a politically charged identifier intended to frame the subjects negatively for certain audiences — for example: labeling candidates as socialists to trigger ideological opposition
“Democratic socialists”
Appeal To Fear 70%
The framing implies Trump is reintroducing a feared ideological threat into American politics, designed to provoke anxiety — for example: suggesting the return of 'communists' as a danger to the country
“bringing back "communists"”
Analyzed automatically with Biasly'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 →