Search results for Prompt Injection Protection
AI & Enterprise
Snowflake expands security push, says agentic AI security needs fundamentally different approach
Snowflake unveiled a slate of new security products at its Snowflake Summit 26 annual conference in San Francisco, focusing on a security model for the agentic era and implementing zero-trust security for AI agents. The company introduced agent identity management, AI security posture management, ransomware and data exfiltration prevention, and prompt injection protection. It framed its strategy around secure guardrails, centralised governance, and data and AI protection.
AI & Enterprise
Snowflake upgrades Horizon Catalog, integrating governance, business context and security
Snowflake said on June 2 it updated its Horizon Catalog data governance platform, focusing on integrating AI agent governance, data contextualisation and security. The company highlighted Horizon Context, designed to unify business logic across an enterprise so tools, teams and AI agents operate on the same data definitions. Snowflake also strengthened security features, including agent identity verification, monitoring via a Trust Center, and added protections against jailbreak attempts and zero-day vulnerabilities. It also introduced Adaptive Compute to optimise resources for AI workloads.
AI & Enterprise
Microsoft rolls out major security update for AI agents
Microsoft unveiled a broad set of security updates aimed at addressing the spread of AI agents. The update covers agent governance, identity protection, data security, cloud and endpoint defense, and AI-based security operations. A key element is Microsoft Agent 365, due for general release on May 1, which centralises management of AI agents. The company also introduced new visibility tools, expanded identity features, embedded data controls into AI workflows, and accelerated security operations automation.