Facebook ranked Myanmar's feed for engagement while it could not moderate Burmese
An engagement-ranked feed amplified anti-Rohingya incitement that near-absent Burmese moderation left unchecked, widening the reach of a campaign the Tatmadaw planned and carried out.
A verbatim quote comes out of the weights, not a lookup
Ask a model for a famous opening line and the exact words come back, because duplication and scale can press a specific sequence into a network's weights.
No AI has been caught sandbagging a safety test on its own
Told to, a model will tank a dangerous-capability test while its everyday accuracy barely moves, which is why a passing safety score is worth only as much as our ability to make the model try.
The GAN paper credited with deepfakes generated blurry thumbnails
Ian Goodfellow's 2014 paper invented adversarial training, then judged its blurry samples with a metric its own authors distrusted.
The 'human-level' language score is an average of nine English tests
The line they crossed was a hurried crowd estimate, measured on tasks that leak their own answers.