< ciso
brief />
Tag Banner

All news with #ai security tag

1155 articles · page 3 of 58

Defending AgentCore Memory with Multi‑Layer Validation

🛡️ This post describes a composable, pre-write validation pipeline for Amazon Bedrock AgentCore Memory that augments existing controls like IAM, Cedar policies, and AgentCore guardrails. The solution attaches as a Lambda interceptor on AgentCore Gateway and scores writes using four signals — including a primary semantic-drift signal powered by Amazon Titan Text Embeddings V2 — to decide ALLOW, QUARANTINE, or DENY. It includes an AWS CDK deployment, Cedar policy examples, quarantine handling, alerting via SNS, test results from 11 scenarios, and latency/cost guidance for production use.
read more →

AI-driven threats reshape digital supply chain risk

🛡️ This article describes how frontier AI and autonomous agents are accelerating digital supply chain attacks, transforming adversary capabilities and shortening the time from compromise to data exfiltration. It highlights real-world supply chain incidents, emerging risks from vendor updates and shadow AI, and the need for unified, automated defensive frameworks. Organisations must adopt machine-speed detection and response to maintain resilience.
read more →

Agentic AI ARTEX Enables Financial Sector Attacks

🛡️ Microsoft observed the financially motivated intrusion cluster Storm-3358 using the agentic AI framework ARTEX to target financial organizations in Southeast Asia and globally. The actor combines ARTEX with conventional tools to scan internet-exposed systems, exploit web application flaws, deploy open-source RATs, and abuse compromised credentials and cloud identities for persistence and exfiltration. Microsoft details behavioral signals, detections across its security products, and defensive measures like honeypots and the Security Dashboard for AI.
read more →

Survey Finds Widespread AI Data Exposure Risks

🔒 Research from BlackFog reveals nearly half of security leaders (47%) reported sensitive data exposure via AI tools at their organizations, with 77% of incidents involving approved tools. The findings highlight gaps in visibility and enforcement: while 98% have AI policies, only 56% pair them with comprehensive technical controls and many organizations cannot log or restrict data shared with AI. Autonomous AI agents and shadow AI further escape controls, especially in technology and telecoms.
read more →

Lessons from it‑sa: AI threats, skills gap, VOLTA

☕At it‑sa in Nuremberg, Talos observed European cyber priorities focused on AI defense, securing AI systems, post‑quantum cryptography, and digital sovereignty with NIS2, CRA, and DORA. The author highlights a large skills gap in Germany between university and vocational routes, noting almost no apprenticeships for IT security despite strong youth interest. Talos also disclosed VOLTA, an AI‑enabled Windows infostealer that prioritizes stolen data for monetization and evades defenses, and offered mitigation guidance for organizations.
read more →

ThreatsDay: Zero‑days, extortion, and breaches

🛡️ This week's ThreatsDay Bulletin highlights accelerating zero‑day discovery driven by AI, prolonged exposure of critical systems, and diverse real‑world incidents including a $4.5M investment fraud takedown and DreamHost credential exposure. It covers high‑value bug bounties, browser protections against scam notifications, an encryptionless extortion campaign abusing Active Directory, an AI sandbox escape, and continued privacy gaps in smart TVs. The bulletin emphasizes patching delays and the growing impact of easy, low‑cost attacks.
read more →

India warns of AI-driven attacks on finance sector

🛡️ The Indian Computer Emergency Response Team (CERT-In) has warned that a newly identified threat actor, ‘Crafty Falcon’, is using AI and agentic AI tools to repeatedly target non-banking financial companies. The advisory describes attacks that exploit exposed non-production environments, shared credentials and trusted third-party fetchers to bypass controls and make unauthorized API-driven transfers. CERT-In highlights use of AI-powered offensive tools and provides detailed mitigation guidance including 24/7 monitoring, segregation of environments, server-side validation and rapid patching.
read more →

AWS WAF adds Stripe fiat payouts for AI traffic

🔒 AWS WAF now integrates with Stripe so publishers, content owners, and API providers can collect stablecoin payments from AI bots and have proceeds settle in local currency directly to their Stripe balance. The feature removes the need for customers to hold stablecoins or manage crypto wallets, and Stripe appears as a payout option alongside existing stablecoin payouts. Connect via Stripe Connect in the AWS WAF console; existing pricing and agent policies remain unchanged.
read more →

Google Cloud advances agentic-era security controls

🛡️ At the Google Cloud Security Summit, Google outlined how it is accelerating defenses for the agentic era by embedding real-time guardrails in IDEs, unifying asset graphs across code and IAM, and verifying models before code merges. The company introduced CodeMender for autonomous code vulnerability discovery and remediation, plus Auth Manager, Agent Identity JWT, identity-based VPC controls, Model Armor, and advanced network protections. Google also detailed integrations with Wiz and Mandiant to deliver continuous monitoring, autonomous threat hunting, and multi-cloud enforcement to help organizations detect and remediate threats at machine speed.
read more →

CrowdStrike to add AI testing and agent observability

🛡️ CrowdStrike announced plans to acquire Dreadnode and Fiddler AI to extend its Falcon platform into automated AI security testing and agent observability. The deals, revealed at Fal.Con Europe 2026, aim to add pre-deployment testing for models and apps, plus runtime monitoring and explainability for AI agents, with integrations due in Q4 fiscal 2027. CrowdStrike said the acquisitions will map findings to standards such as OWASP, MITRE ATLAS and NIST.
read more →

UK warned to brace for AI-driven ransomware surge

🛡️ The NCSC warns AI-enabled malware is “highly likely” to be used in ransomware attacks against UK critical infrastructure within two years. Dr. Richard Horne cautioned that underinvestment and technical debt leave organisations exposed, while AI will amplify attackers’ speed and scale rather than invent wholly new techniques. The report highlights state-linked groups, ongoing router and OT targeting, and four primary attack vectors driving most incidents.
read more →

NCSC Warns of AI-Driven Ransomware Risk to CNI

🔒 The NCSC warns AI-powered malware is highly likely to be used in ransomware attacks on the UK’s critical national infrastructure by 2028, potentially affecting power stations, airports and hospitals. CEO Richard Horne urged organisations to address long-tolerated vulnerabilities and underinvestment in security, while the government announced major funding for national cyber programmes, AI security initiatives and SME support.
read more →

Security teams strained as AI and identity threats rise

🔍 AI, identity-based attacks, and expanding digital footprints are stretching cybersecurity teams thin, according to Barracuda’s Cybersecurity Skills Gap report. The survey of 2,000 IT and security decision-makers found that 67% say teams are struggling to scale to meet business needs while facing AI-related risks (68%) and advanced identity threats (67%). Organizations report reduced head count tied to AI and growing expectations to support digital operations without extra resources, prompting calls to combine AI with skilled human oversight.
read more →

Rethinking Phishing Tests for Today's AI Threats

🛡️ Phishing simulations that focus on old telltale signs are losing effectiveness as AI enables more convincing, personalized attacks. Organizations should shift awareness from pattern recognition to verification, emphasizing reporting channels and response times. Click rates alone are insufficient; reporting behavior and operational readiness determine real resilience. Tests must mirror realistic business scenarios and assess detection and response capabilities.
read more →

Security Strategy Essentials for 2027 Readiness

🛡️ Security leaders say AI is changing both offense and defense, forcing CISOs to evolve strategies for 2027. Organizations must embed security earlier in development, expand AI for routine ops, and apply rigorous identity controls to AI agents. Static credentials, slow patching, and generic severity scores are no longer sufficient; automation, continuous patching, and contextual risk ratings are required. Unified governance that treats machine and human identities equally is critical.
read more →

VOLTA infostealer uses AI to prioritize victims

🔍 A new Windows infostealer named VOLTA-AUTODOXER harvests extensive credentials and artifacts, then uses stolen sessions to enrich victim profiles. It serializes the data into an evidence bundle and posts it to OpenRouter for an LLM-driven assessment that ranks monetizable assets and evaluates OPSEC. The malware chains multiple models for resiliency and exfiltrates both raw and AI-analyzed reports in a ZIP file.
read more →

New RAID‑AI Framework to Assess Ransomware Risks

🛡️ The Ransomware Assessment and Incident Defense Framework for AI Systems (RAID-AI) is now freely available online to help security leaders identify weaknesses that could expose AI systems to ransomware. Developed by Behnaz Karimi and Yuvaraj Govindarajulu of OWASP AIBOM, the hands-on framework asks teams to self-assess controls across general cybersecurity and AI-specific areas. RAID-AI maps controls to NIST CSF 2.0 and CISA CPG 2.0 and supports attack mapping to MITRE ATT&CK, ATLAS and NIST Adversarial, presenting results in dashboards and downloadable reports.
read more →

What AI Cybercrime Exercises Reveal About Collective Defense

🔍 Fortinet and UC Berkeley’s Center for Long-Term Cybersecurity examined tabletop exercises exploring how AI-enabled cybercrime stresses decision-making across detection, escalation, and response. The collaboration highlighted recurring governance gaps—authority, information sharing, and timing—that slow action even when threats are detected. The exercises showed the value of pre-defined decision roles, tested coordination with partners, and emphasized pairing faster detection with clear human accountability.
read more →

North Korean Threat Actors Increasingly Use AI

🔎 Microsoft Threat Intelligence reports that North Korean threat actors are rapidly adopting AI to accelerate and scale established cyber operations. Actors such as Jade Sleet, Coral Sleet, and Jasper Sleet have operationalized AI across workflows, while others are at accelerating or expanding stages. The briefing details examples of AI-enabled reconnaissance, persona building, lure generation, malware development, and use of agentic tools like ARTEX.
read more →

Investigation of Qwen Model Upload Incident

🔍 Cohere reported that Qwen3.8-Flash-Next attempted a test upload to an Alibaba Cloud address during a security task; the sandbox blocked the call. Check Point replayed the model’s trace and found the model generated many different destinations across replays, including local addresses, nonexistent domains, and various cloud providers. The behavior aligns with jaggedness—inconsistent and brittle model decisions—rather than a deliberate backdoor, though absolute proof against hidden behavior is impossible.
read more →