Tag: Security
All the articles with the tag "Security".
The Middle East's AI Arms Race: Experimentation and Escalation
Published: at 10:33 AMThe Middle East is becoming a hub for AI-driven military technologies, raising concerns about escalating conflicts, increasing civilian casualties, and the lack of ethical oversight in the region's arms race.
Microsoft Pauses Recall Feature Rollout for AI PCs Following Security Concerns
Published: at 04:03 AMMicrosoft delayed the launch of its "Recall" AI PC feature due to significant privacy and security concerns, opting for a limited release to Windows Insiders for further testing and refinement before a broader rollout.
Chrome Extension Using AI Engine Acted as Man-in-the-Middle Proxy Stealing User Data
Published: at 05:35 PMA malicious Chrome extension disguised as an AI writing assistant acted as a Man-in-the-Middle proxy, stealing user credentials and other sensitive data. Prompt action by researchers and Google prevented further harm.
Social Security's AI Training Video Stumbles on Accessibility and Accuracy
Published: at 02:09 PMSocial Security's AI training video has sparked criticism for factual inaccuracies and a lack of accessibility features, raising concerns about responsible AI usage and potential harm to vulnerable populations.
All Major GenAI Models Vulnerable to 'Policy Puppetry' Prompt Injection Attack
Published: at 10:13 AMThe "policy puppetry" attack exposes a critical vulnerability in major GenAI models, enabling malicious actors to bypass safety measures through subtle prompt manipulation. This highlights the need for improved security and contextual understanding within these systems.
AI's Growing Importance in U.S. National Security: A Strategic Overview
Published: at 03:10 AMThe article details the US Department of Defense's strategic adoption of AI to enhance national security, focusing on situational awareness, decision-making, automation, and cyber defense, while emphasizing responsible development through collaboration.
Microsoft's Taxonomy of Failure Modes in AI Agents: A Security Perspective
Published: at 09:55 PMMicrosoft's whitepaper introduces a detailed taxonomy of failure modes in AI agents, categorizing potential risks into seven classes to guide developers and researchers in building safer, more reliable, and ethically sound AI systems.
Meta's WhatsApp Integration of AI Chatbot Sparks Privacy Concerns and User Outrage
Published: at 02:14 PMMeta's automatic integration of an AI chatbot into WhatsApp has triggered user outrage due to privacy concerns and a lack of opt-out options. The move is perceived as intrusive and raises questions about data collection.
Protect AI and Leidos Partner to Secure AI Across U.S. Government Systems
Published: at 02:09 PMProtect AI and Leidos are partnering to secure AI in the U.S. government, integrating Protect AI's platform into Leidos' offerings to protect against AI-specific vulnerabilities.
Endor Labs Deploys AI Agents to Mitigate Supply Chain Coding Risks
Published: at 04:29 PMEndor Labs is using AI agents to automate the identification and remediation of security risks in open-source dependencies, combating "vibe coding" and empowering developers with secure alternatives.