AI Briefing#model-security
The latest AI and developer news about #model-security, with the original source and a short summary.
The latest AI and developer news about #model-security, with the original source and a short summary.
The study demonstrates that Qwen3-Coder generates less secure code for US and Hong Kong personas due to weight-level alignment, a bias that persists despite security prompts but is removed by a 'Westernized' weight modification.
New research shows that while explicit hidden prompts are ineffective and sanctionable, subtle preference framing significantly alters AI legal analysis outcomes.
DLR-Lock makes unauthorized modification of pre-trained weights difficult while maintaining model performance.
RiskRubric.ai has launched, evaluating AI models' security, privacy, safety, and more using standardized metrics.
Hugging Face is adopting JFrog's scanner to enhance detection of malicious code within model weights.
Hugging Face is enhancing its model security scanning capabilities by integrating Protect AI's Guardian.