Reveals fundamental limitations in current AI safety evaluation methods for researchers and security experts.
Researchers prove AI models can hide arbitrary capabilities using cryptographic backdoors that resist elicitation through and adversarial methods.
Researchers prove AI models can cryptographically hide arbitrary capabilities using backdoor attacks that embed encrypted neural networks only triggered by secret inputs.
OpenAI claims a faster Fourier transform algorithm, but analysis suggests it's more exploit than breakthrough with limited practical benefits.
Anthropic research shows small AI models may soon self-replicate on consumer IoT devices, potentially creating unstoppable malware that adapts across different hardware.
Harvard study finds AI coding agents increase code output by 30% but create review bottlenecks, resulting in no net gain in software production or employment reduction.
Researchers developed Prompteina, the first all-atom protein structure generation model that accepts natural language descriptions as input.
