All news with #ai security tag

The AI Fix 74: AI Glasses, Deepfakes, and AGI Debate

🎧 In episode 74 of The AI Fix, hosts Graham Cluley and Mark Stockley survey recent AI developments including Amazon’s experimental delivery glasses, Channel 4’s AI presenter, and reports of LLM “brain rot.” They examine practical security risks — such as malicious browser extensions spoofing AI sidebars and AI browsers being tricked into purchases — alongside wider societal debates. The episode also highlights public calls to pause work on super-intelligence and explores what AGI really means.

Building Data Security from the Inside Out: Hybrid Focus

🛡️ Cybersecurity Awareness Month underscores that protecting organizational data requires attention to internal handling as well as external threats. Fortinet’s 2025 Insider Risk Report found 77% of organizations experienced insider-related data loss in the past 18 months, with nearly half of incidents tied to simple negligence. The report highlights mounting GenAI concerns and recommends a layered approach combining visibility, behavioral analytics, and real-time coaching to prevent accidental and malicious loss.

AI-Powered, Quantum-Ready Network Security Platform

🔒 Palo Alto Networks presents a unified, AI-driven approach to network security that consolidates browser, AI, and quantum defenses into the Strata Network Security Platform. New offerings include Prisma Browser, a SASE-native secure browser that blocks evasive attacks and brings LLM-augmented data classification to the endpoint, and Prisma AIRS 2.0, a full-lifecycle AI security platform. The company also outlines a pragmatic path to quantum-readiness and centralizes control with Strata Cloud Manager to simplify operations across hybrid environments.

Prisma AIRS 2.0: Unified Platform for Secure AI Agents

🔒 Prisma AIRS 2.0 is a unified AI security platform that delivers end-to-end visibility, risk assessment and automated defenses across agents, models and development pipelines. It consolidates Protect AI capabilities to provide posture and runtime protections for AI agents, model scanning and API-first controls for MLOps. The platform also offers continuous, autonomous red teaming and a managed MCP Server to embed threat detection into workflows.

Cortex AgentiX: Agentic AI Platform for Autonomous SOC

🤖 Palo Alto Networks introduces Cortex AgentiX, an agentic AI platform designed to build, deploy and govern autonomous security and IT agents. The vendor says AgentiX extends the Cortex foundation and leverages 1.2 billion playbook executions to deliver end-to-end agentic workflows and drive up to a 98% reduction in Mean Time to Respond with 75% less manual work. It ships with prebuilt agents for threat intelligence, email, endpoint, network, cloud and IT, and highlights full transparency, role-based controls and human-in-the-loop approvals. AgentiX is embedded in Cortex XSIAM and Cortex Cloud today; a standalone platform and Cortex XDR integration are slated for early 2026.

Quarter of Scam Victims Report Considering Self-Harm

⚠️ A new 2025 Consumer Impact Report from the Identity Theft Resource Center (ITRC) finds identity fraud is driving severe mental and financial harm, with one quarter of surveyed consumers saying they seriously considered self-harm after an incident. The figure rises to 68% among self-identified victims but falls to 14% for those who contacted the ITRC, underscoring the value of professional support. The study of 1,033 general consumers also highlights rising repeat victimisation, large monetary losses — including more than 20% losing over $100,000 and 10% losing at least $1m — social media account takeovers as the most common crime, and widespread concern that AI will be a major battleground for identity security.

Copilot Mermaid Diagrams Could Exfiltrate Enterprise Emails

🔐 Microsoft has patched an indirect prompt injection vulnerability in Microsoft 365 Copilot that could have been exploited to exfiltrate recent enterprise emails via clickable Mermaid diagrams. Researcher Adam Logue demonstrated a multi-stage attack using Office documents containing hidden white-text instructions that caused Copilot to invoke an internal search-enterprise_emails tool. The assistant encoded retrieved emails into hex, embedded them in Mermaid output styled as a login button, and added an attacker-controlled hyperlink. Microsoft mitigated the risk by disabling interactive hyperlinks in Mermaid diagrams within Copilot chats.

Atlas Browser Flaw Lets Attackers Poison ChatGPT Memory

⚠️ Researchers at LayerX Security disclosed a vulnerability in OpenAI’s Atlas browser that allows attackers to inject hidden instructions into a user’s ChatGPT memory via a CSRF-style flow. An attacker lures a logged-in user to a malicious page, leverages existing authentication, and taints the account-level memory so subsequent prompts can trigger malicious behavior. LayerX reported the issue to OpenAI and advised enterprises to restrict Atlas use and monitor AI-driven anomalies. Detection relies on behavioral indicators rather than traditional malware artifacts.

Recruitment red flags: spotting faux job applicants

🔍 Organizations are facing a growing threat from applicants who pose as legitimate job seekers but are in fact operatives tied to overseas actor networks. Recent cases — including a July 2024 incident at KnowBe4 and longer running campaigns tracked as WageMole and DeceptiveDevelopment — show perpetrators use stolen identities, deepfakes and remote infrastructure to gain employment. The article outlines practical detection cues for recruitment teams and containment steps to limit insider risk.

Rethinking Service Provider Risk: A CISO Imperative

🔍 As organizations outsource more critical systems and security functions to managed service providers, the complexity and frequency of third-party incidents are rising — 47% of organizations reported a third-party breach in the 12 months to mid-2025. Security leaders must balance rigorous, standards-based assurance (for example ISO 27001 or SOC 2) with relationship-driven vetting that fosters transparency and shared responsibility. Experts from media company Advance, the University of Queensland and vendor advisors argue that questionnaires alone are insufficient: meaningful dialogue, selective disclosure (summaries of pen tests rather than full reports), contractual clarity, and AI-aware controls are all needed to assess and manage evolving risks.

ChatGPT Atlas 'Tainted Memories' CSRF Risk Exposes Accounts

⚠️ Researchers disclosed a CSRF-based vulnerability in ChatGPT Atlas that can inject malicious instructions into the assistant's persistent memory, potentially enabling arbitrary code execution, account takeover, or malware deployment. LayerX warns that corrupted memories persist across devices and sessions until manually deleted and that Atlas' anti-phishing defenses lag mainstream browsers. The flaw converts a convenience feature into a persistent attack vector that can be invoked during normal prompts.

Exposure Management in 2025: Trends, Risks, and Response

🔒 Intruder’s 2025 Exposure Management Index analyzes scans from over 3,000 small and midsize businesses to show defenders adapting under mounting pressure. High-severity vulnerabilities rose nearly 20% year‑on‑year, even as 89% of resolved critical flaws were remediated within 30 days (up from 75% in 2024). The report highlights AI-driven exploit development, growing attack surfaces from cloud, shadow IT and supply‑chain risk, and faster remediation at smaller firms.

Cloudflare Radar's Evolution: Expanding Internet Observability

📡 Since its 2020 debut, Cloudflare Radar has evolved into a comprehensive observability platform that aggregates Cloudflare telemetry to illuminate security, performance, and usage trends. Initially centered on Radar Internet Insights, Domain Insights, and IP Insights, the service has grown to include Certificate Transparency metrics, TCP reset/timeouts visibility, post-quantum adoption tracking, and AI-focused crawler analytics. Radar also added routing tools such as route leak and origin hijack detection, real-time BGP views, AS-SET monitoring, and notifications, while improving programmatic access via the Radar API and an MCP server for LLM integration. Popular utilities like the URL Scanner, expanded search and date-range options, and internationalized interfaces reinforce Radar's mission to make the Internet more observable and resilient.

Internet Measurement, Resilience and Transparency Week

📡 This week Cloudflare Research publishes a series of posts revealing methods and findings that advance a more measurable, resilient, and transparent Internet. The series explores Internet measurement fundamentals, resilience frameworks, post-quantum deployment, and networking innovations, with deep dives into products such as Cloudflare Radar and experiments like Merkle Tree Certificates. Expect practical analysis, IETF-aligned protocol discussion, and real-world deployment considerations.

Proving Data Sovereignty: Controls, Keys, and Audits

🔒 The article argues that data sovereignty commitments like Project Texas must be supported by auditable, technical evidence rather than marketing promises. It prescribes five concrete, testable controls — brokered zero‑trust access, in‑region HSM keys, immutable WORM logs, continuous validation, and third‑party attestation — plus measurable metrics to prove compliance. A 90‑day blueprint and emerging AI automation are offered to operationalize verification and produce regulator‑ready, reproducible evidence.

How MDR Gives MSPs a Competitive Market Edge Today

🛡️ Managed detection and response (MDR) helps managed service providers (MSPs) overcome talent shortages, alert overload and rapidly evolving threats by outsourcing 24/7 SOC monitoring, behavioral detection, threat hunting and automated incident response. MDR can open recurring revenue streams, strengthen customer relationships and meet cyberinsurance conditions, while intelligent prioritization and GenAI-assisted playbooks reduce operational strain and false positives. Choosing a partner with proven threat intelligence, continuous operations and a human-plus-machine approach is critical.

OpenAI Atlas Omnibox Vulnerable to Prompt-Injection

⚠️ OpenAI's new Atlas browser is vulnerable to a prompt-injection jailbreak that disguises malicious instructions as URL-like strings, causing the omnibox to execute hidden commands. NeuralTrust demonstrated how malformed inputs that resemble URLs can bypass URL validation and be handled as trusted user prompts, enabling redirection, data exfiltration, or unauthorized tool actions on linked services. Mitigations include stricter URL canonicalization, treating unvalidated omnibox input as untrusted, additional runtime checks before tool execution, and explicit user confirmations for sensitive actions.

Top 10 Challenges Facing CISOs and Security Teams Today

🔒 Security leaders face a rapidly evolving threat landscape driven by AI, constrained budgets, talent shortages, and a vastly expanded attack surface. Many organizations rushed into AI adoption before security controls matured, and CISOs report growing involvement in AI governance and implementation even while attackers leverage AI to compress time-to-compromise. Data protection, employee susceptibility to sophisticated scams, quantum readiness, and board alignment emerge as immediate priorities that require clearer risk-based decisions and frequent simulation exercises.

AI 2030: The Coming Era of Autonomous Cybercrime Threats

🔒 Organizations worldwide are rapidly adopting AI across enterprises, delivering efficiency gains while introducing new security risks. Cybersecurity is at a turning point where AI fights AI, and today's phishing and deepfakes are precursors to autonomous, self‑optimizing AI threat actors that can plan, execute, and refine attacks with minimal human oversight. In September 2025, Check Point Research found that 1 in 54 GenAI prompts from enterprise networks posed a high risk of sensitive-data exposure, underscoring the urgent need to harden defenses and govern model use.

Cyber-risk in the Shadows: Shadow IT, AI Use and Risks

🛡️ In a short video for Cybersecurity Awareness Month, ESET Chief Security Evangelist Tony Anscombe explains how unsanctioned hardware and software — commonly called shadow IT — is creating security gaps in the remote and hybrid work era. He warns that growing employee use of generative AI further increases risk by exposing sensitive corporate data outside IT control. The video outlines practical steps IT teams can take to discover, govern and mitigate these hidden risks and points to related guidance on authentication, patching and ransomware resilience.