NVIDIA and OSAI unveil SAFE framework for AI incident response
After a Hugging Face breach incident, NVIDIA and the Open Secure AI Alliance (OSAA) introduced the SAFE framework to help keep AI safer.
The idea is to make it easier for companies to share info about hacks, respond in a more organized way, and cut down on future risks.
SAFE proposals open for public feedback
SAFE's proposals were developed during discussions at the Black Hat cybersecurity conference in Las Vegas and have been opened for public feedback.
It focuses on private reporting of AI security issues, notifications for affected parties, and teamwork after incidents happen.
The Linux Foundation is managing the review process for the proposals, which are open for public feedback.
More than 120 companies have already joined, including Microsoft, Intel, Adobe, and Visa, but big players like OpenAI and Google have not signed up yet.