Google, OpenAI, Microsoft, Amazon, Nvidia, Intel and other big names in AI are coming together to form the Coalition for Secure AI (CoSAI), according to a announcement on Thursday. The initiative aims to address a “fragmented AI security landscape” by providing access to open source methodologies, frameworks, and tools.
Other companies that have joined CoSAI include IBM, PayPal, Cisco and Anthropic. CoSAI will exist within the Organization for the Advancement of Structured Information Standards (OASIS), a non-profit group that promotes the development of open standards. To begin, CoSAI will work toward three goals: developing best practices for AI security; face challenges in AI; and secure AI applications.
“We have used AI for many years and see the continued potential for defenders, but we also recognize its opportunities for adversaries,” Heather Adkins, vice president of security at Google, said in a statement. “CoSAI will help organizations, large and small, integrate AI safely and responsibly – helping them realize its benefits while mitigating risks.”