NvidiaのSAFE、OpenAIやGoogle不在のままAIセキュリティの脅威に対抗
Nvidiaが先週設立したOpen Secure AI Allianceは、インシデント報告の標準化とセキュリティツールのカタログ化を目的にSAFEを立ち上げた。OpenAIとGoogleが不参加の中、企業向けAIの防御を目指す。
nvidia's open secure ai alliance launched the shared ai findings exchange to standardize reporting of ai cybersecurity incidents and catalog defenses like the garak scanner. the group of 120 members, including microsoft and amazon, is rapidly forming protocols to secure ai agents and open source models. notable absences include openai and google, despite their prior support for open source advocacy. the move signals a rapid industry mobilization to secure the ai ecosystem against threats like rogue models, potentially spurred by recent u.s. policy concerns regarding chinese open weight ai.