All news with #ai security tag

🔒 As AI models like GPT5.5 and Claude Mythos accelerate exploit discovery, organisations face shrinking windows to patch vulnerabilities. Industry experts at Infosecurity Europe warn mean time to exploit has fallen from days to hours, prompting regulatory responses such as India’s 12-hour patch expectation. Analysts contrast vendor-centric EU rules with market-driven US approaches and recommend exploit-intelligence led patching, automation, segmentation and stronger producer SLAs.

🛡️ This article explains why AI requires a unified security architecture — the AI Defense Plane — to discover, protect, govern, and assure AI behavior across employees, applications, and agents. It describes how AI becomes an execution layer that can retrieve data, call tools, and take actions, creating risks that traverse traditional security boundaries. The piece emphasizes runtime protection, coordinated enforcement, and continuous testing to prevent prompt-based attacks, data exposure, and unsafe agent behavior.

🔍 Halcyon research reveals a dramatic rise in AI-powered cybercrime tooling across underground markets, jumping from 38 mentions in December to 1,486 in February. Cynthia Kaiser, SVP of Halcyon’s Ransomware Research Center, detailed four product categories: weaponized LLMs, AI-enabled identity fraud, AI-augmented malware/infrastructure, and jailbroken or stolen AI services. She warned that automated distribution, freemium models and redundant channels have lowered the financial barrier to entry and increased resilience against takedown efforts.

🔒 Reporting suggests attackers used Meta’s AI support chatbot to change recovery emails on high-profile Instagram accounts, leading to notable takeovers. The core issue isn’t just prompt injection or a model jailbreak but that the AI operated within a sensitive account recovery workflow with insufficient independent verification. Organizations must treat AI-driven support actions as part of the security boundary and constrain authority, permissions, and verification around such agents.

🧭 Microsoft outlines a shift from isolated AI tools to a unified, enterprise-grade agent platform that runs real work. The post emphasizes a single integrated system spanning Azure, GitHub, Microsoft IQ, Foundry, Agent 365, and Microsoft 365 to build, contextualize, run, govern, and improve agents. It stresses secure-by-design governance, model choice, continuous improvement through feedback and tuning, and production-grade runtimes. The approach centers developers and enterprise context to make agents trustworthy and scalable.

🔎 Foundry IQ streamlines bringing enterprise and external knowledge into agent workflows by unifying content, improving ingestion, and offering a serverless model for retrieval. The service provides a Model Context Protocol (MCP) server, integrates Microsoft Web IQ for low-latency external context, and includes GA security and compliance features. Serverless Developer tier is in public preview with CU-based billing estimates and scale-to-zero capacity.

🛡️ Joe Slowik warned at Infosecurity Europe that defenders must adopt AI to keep pace with adversaries. He argued that purely human-driven SOCs cannot match the accelerated timescales enabled by AI, ML and LLMs, leaving organisations exposed. Slowik recommended rethinking security operations to integrate AI agents for rapid intelligence, enrichment and remediation, while keeping humans in the decision loop. He used the React2Shell example to illustrate the speed of modern exploits.

🔒 Bruce Schneier reflects on his 2010 Dark Reading essay arguing that while cryptography provides strong mathematical advantages, it cannot by itself secure modern, interconnected systems. He traces how crypto has been applied since the 1990s and explains that computer security is an ongoing arms race of fragile defenses. Schneier warns that AI changes the landscape by automating vulnerability discovery and exploit creation, shifting the balance between attackers and defenders.

🔍 Sophos X-Ops uncovered a lab where a threat actor used AI coding tools to develop and test malware aimed at evading EDR products. The files and Git repository showed Python scripts—many partially AI-generated—used to build and iterate evasion modules against vendors including Sophos, CrowdStrike and Microsoft. Humans retained control of the workflow, using AI to accelerate building, testing and refinement while operating inside an AI-native environment.

🔍 New research from ManageEngine reveals UK IT and business leaders view AI-powered cyber-attacks as their top risk over the next 12 months, with 43% identifying it as the single biggest threat. The survey of 1,500 decision-makers across five European markets shows 41% of UK respondents plan to prioritise spending on tackling AI and advanced threats. Despite strong detection rates, UK organisations report increasing incidents, skills gaps and recovery challenges, alongside rising investment in resilience and governance.

🛡️ Responsible Disclosure in the Age of AI argues that frontier AI systems now autonomously discover software vulnerabilities at unprecedented speed and scale, exposing long-standing technical debt in the software industry. The piece traces the evolution of assurance practices and disclosure frameworks and highlights growing tension between offensive and defensive cyber equities, particularly in the U.S. and China. It calls for coordinated national and international efforts to accelerate remediation, patch management, and investment in automated repair capabilities to close the narrowing window before adversaries exploit these advances.

🔧 AWS Parallel Computing Service (AWS PCS) now offers a PCS‑ready Deep Learning AMI, an AWS‑maintained Amazon Machine Image based on the Deep Learning Base GPU AMI (Ubuntu 24.04). It provides a production‑quality foundation for AI/ML training and HPC with preinstalled, compatibility‑tested infrastructure components such as NVIDIA drivers, CUDA, EFA, Lustre client, PCS Agent, Slurm for PCS, and EFS utilities. Multiple Slurm versions are supported and activate automatically based on cluster configuration, and AWS will regularly update the AMIs for security patches and driver updates. The AMI is available at no additional cost for x86_64 and arm64 in all Regions where AWS PCS is offered.

🛡️ The Remote Model Context Protocol (MCP) Server for AlloyDB is now generally available, providing a secure HTTP endpoint that lets AI agents access real-time operational data. This fully managed service simplifies production deployments by centralizing discovery, offering fine-grained IAM-based authorization, audit logging, and integration with Model Armor for prompt and response protection. Developers can join AlloyDB operational data with analytics in BigQuery and use built-in AI functions for low-latency agentic experiences.

🛡️ Vendors at Infosecurity Europe 2026 agree that AI will not eliminate security operations centers but will automate repetitive triage and ticketing. Experts urge treating AI as a glass box, ensuring transparency and human-in-the-loop validation. The shift accelerates junior analysts into supervisory tier-1.5 roles and creates demand for cyber defense engineers who build and tune detection systems.

🧭 At Infosecurity Europe 2026, OWASP will unveil the Agentic Research Council to better align fast‑moving agentic AI capabilities with security research and operational practice. Launched from the GenAI Security Project’s Agentic Security Initiative, the council will prioritize a public research pipeline, convene working groups and connect academic outputs to deployable mitigations. The initiative aims to accelerate runtime‑focused defenses against multi‑agent threats.

📩 Kaspersky examines how everyday messaging channels like SMS, WhatsApp, and email are being exploited by organized scam cartels that use speed, familiarity, and AI to trick victims. The research shows average losses of $733 per victim, rapid attack timelines often under 30 minutes, and widespread emotional damage eroding trust in digital communications. The post highlights common schemes, platform distribution, and recommendations to protect yourself.

🔒 CISOs admit many organizations remain underprotected, with surveys showing gaps in data protection, incident preparedness, and resourcing. As adversaries adopt automation and AI, security programs must close six core gaps: perception, speed versus attackers, business‑security alignment, skills, AI security, and legacy systems. Experts urge CISOs to shift toward resilience, accelerate operations with automation and CTEM, and invest in workforce and governance.

🔒 At GTC Taipei during COMPUTEX 2026, NVIDIA highlighted its Vera BlueField-4 STX and DOCA innovations designed to secure enterprise AI infrastructure. Modern AI factories combine high-performance compute, distributed storage, Kubernetes, APIs, GPU farms, and sensitive data, creating new security needs. Check Point integrates its AI Factory Firewall with NVIDIA BlueField and DOCA to provide visibility, segmentation, runtime protections, and infrastructure-level policy enforcement across distributed AI environments.

🛡️ This Cloud CISO Perspectives post by Usman Chaudhary, Field CISO for Google Public Sector, outlines a pragmatic roadmap for public-sector CISOs to adopt AI-driven security. It emphasizes immediate quick wins in the first 90 days, tactical actions within six months, and strategic initiatives for months six to 12, combining internal automation, commercial AI capabilities, and vendor solutions like Gemini for Government. The guidance targets threat triage, talent augmentation, posture elevation, and governance to reduce toil and accelerate proactive defense.

🎥 ESET Chief Security Evangelist Tony Anscombe reviews major cybersecurity stories from May 2026, focusing on industrial control system intrusions, an AI-directed data theft, a Google-reported AI-developed zero-day, and crypto kiosk scams. He outlines attack vectors such as weak passwords and internet-exposed systems, notes the partial failure of an IT-to-OT escalation, and previews mitigation advice for defenders. Watch Tony’s video for practical recommendations and refer to the April edition for additional context.