Prisma AIRS options AI mannequin scanning, which lets enterprises safely undertake AI fashions by scanning them for vulnerabilities and safe the AI ecosystem towards dangers reminiscent of mannequin tampering, malicious scripts, and deserialization assaults. Posture administration gives enterprises with perception into their safety posture as associated to the AI ecosystem and exposes dangers reminiscent of extreme permissions, delicate information publicity, platform misconfigurations, and entry misconfigurations, in line with Palo Alto.
“We imagine Palo Alto is extending its platformization technique deeper into AI with its acquisition of Koi as it may supply management and visibility of AI brokers, plug-ins, and nontraditional software program which have privileged entry talents on an endpoint. In our view, this deal builds on Palo Alto’s current acquisition of Chronosphere within the observability area because it permits the corporate to pair richer AI information with new controls,” wrote Jonathan Ho, a analysis analyst with William Blair Fairness Analysis, in a report on the deal.
“We imagine this could assist Palo Alto higher safe the lifecycle round AI from infrastructure and information to brokers and endpoints, and we view this deal as the most recent in Palo Alto’s strikes to profit from AI spending and safety…it broadens Palo Alto’s protection of dangers round AI on endpoints, which ought to put the corporate in a greater aggressive place for the long run as endpoint safety evolves to incorporate the governance of AI brokers and autonomous workloads on these endpoints,” Ho said. Ho stated Koi’s know-how competes with CrowdStrike, Microsoft, SentinelOne and others.
The Koi deal comes only one week after Palo Alto closed its acquisition of CyberArk, which additionally tackles the safety of enterprise AI belongings. In a blog concerning the CyberArk deal, World Broad Applied sciences said: “It raises the bar for AI safety.
“Each vendor is claiming ‘AI safety.’ Most of it’s detection, guardrails, and posture. However the AI period introduces a brand new id class: agentic processes that may provoke actions, name instruments, use credentials, and persist throughout workflows. If AI brokers are granted privileges, then the query turns into: Who governs the privilege of an agent that by no means sleeps?” WWT said.
“CyberArk has been messaging id safety for AI brokers and non-human identities; Palo Alto is explicitly tying the acquisition to securing “human, machine, and AI identities” at scale. Whether or not the market absolutely agrees but doesn’t matter. This acquisition forces the dialog into the open,” WWT said.
