Many organizations are using artificial intelligence (AI) in various ways, but this integration can expose them to significant security risks, especially when it comes to API vulnerabilities. AI systems often handle sensitive customer data and can unintentionally leak this information if not properly secured. Attackers can exploit weaknesses in these systems, such as through business ...
Databricks has launched important updates to its Mosaic AI Gateway, Unity Catalog tools, and AI/BI Genie, making it easier for organizations to develop reliable AI agents with strong governance and data integration. These updates address significant security and performance challenges by ensuring that AI agents, like those summarizing customer complaints, operate smoothly without risking sensitive ...
Researchers from Michigan State University, the University of Georgia, and Singapore Management University have discovered a new method to manipulate AI models with memory, called MINJA (Memory INJection Attack). Unlike previous threats that required backend access, this attack can be executed simply by interacting with an AI agent like a regular user. This means any ...
Researchers from Michigan State University, the University of Georgia, and Singapore Management University have developed an attack technique called MINJA, which manipulates AI models’ memory through regular user interactions. This is significant because, unlike previous threats that required backend access, anyone can launch MINJA just by using the chatbot. When clever prompts are sent to ...
Varonis Systems has introduced Varonis for Agentforce, enhancing its security capabilities for monitoring AI agents within Salesforce. This new solution gives IT and security teams a clear view of how AI agents connect to vital enterprise data and identifies potential security risks, all from a single dashboard. By building on existing security solutions, Varonis helps ...
Varonis Systems has launched a new solution called Varonis for Agentforce, enhancing its security offerings by providing visibility into the AI layer of Salesforce. This tool allows IT and security teams to monitor AI agents used in various business functions, helping organizations identify potential security risks from a single dashboard. It builds on existing Salesforce ...
Microsoft has announced exciting updates to Azure AI Foundry, the platform for creating and managing advanced AI applications for businesses. The highlight is the introduction of GPT-4.5, offering improved accuracy, reduced errors, and better engagement in tasks like coding and project management. New tools and models, including Phi-4 and solutions from Stability AI, focus on ...
Join the AI Agent & Copilot Podcast as we dive into the latest AI advancements from Microsoft and its partners. This episode features Tom Smith discussing an upcoming session on “AI and Security, Legal Challenges & Policy Considerations” at the AI Agent & Copilot Summit. Guests Nathan Bensch and Corey Hlavacek from enVista share insights ...
As more businesses use autonomous AI agents to automate workflows, traditional single sign-on (SSO) systems struggle to manage non-human identities. This shift requires new authentication methods to handle rapid privilege changes, AI-driven credential management, and the unique speed of AI operations. By 2025, it’s estimated that a significant portion of enterprise workflows will involve AI ...
Zenity, based in Tel Aviv, has been recognized by Gartner as a Representative Vendor in its 2025 Market Guide for Artificial Intelligence Trust, Risk, and Security Management. This acknowledgment highlights Zenity’s innovative approach to securing AI agents by merging AI Security Posture Management and AI Detection Response. Their platform helps businesses identify vulnerabilities in AI ...
