Mistral's Shieldstral 1.0 redefines AI guardrails with policy-adaptive safety
Mistral's Shieldstral 1.0 introduces policy-adaptive AI guardrails, scoring inputs/outputs on a single 16GB GPU, with multimodal support.
Open Secure AI Alliance
Latest Open Secure AI Alliance coverage.
Mistral's Shieldstral: a 3B model that outpaces 7x larger rivals in multimodal safety
Mistral releases Shieldstral, a 3B multimodal safety classifier that outperforms models up to 7x its size, now open-source under Apache 2.0.
Nvidia and Microsoft form Open Secure AI Alliance, excluding OpenAI, Google, and Anthropic
Nvidia and Microsoft launch the Open Secure AI Alliance with 15 other tech companies to develop open-source AI security tools, excluding major players like O...