Search papers, labs, and topics across Lattice.
7
0
9
Scammers are more persistent with older targets, engaging them in longer conversations, yet their requests for sensitive information remain unchanged regardless of age.
Triage performance among on-device models is an illusion; even simple agents can achieve perfect scores, undermining the belief that bigger models always outperform smaller ones.
Twitter strips C2PA provenance data from AI-generated images, making it impossible to cryptographically verify their origin on the platform.
GPT-Image-2 can so seamlessly forge documents that neither humans nor the model itself can reliably tell the difference.
Skip the privacy concerns and annotation costs: this new pipeline lets you generate realistic, labeled eye-tracking data at scale by replaying real gaze trajectories in a 3D simulator.
Existing document forgery detectors are essentially useless against the rising tide of AI-generated document fraud, as revealed by a new benchmark showing a massive performance drop.
A simple synthetic beard fools AI age estimators into classifying minors as adults up to 69% of the time, exposing a critical flaw in current age-verification systems.
