All news in category “AI and Security Pulse”

🔍 Nudge Security provides continuous discovery, monitoring, and governance for shadow AI, delivering a Day One inventory of every AI app and account introduced into an organization. A lightweight IdP integration analyzes machine-generated SaaS emails (without storing content) to detect account creation, password changes, and security setting updates. An optional browser extension monitors AI conversations, flags sensitive data and file uploads, visualizes data flows, and issues real-time nudges and configurable alerts to guide users toward approved tools and enforce acceptable use.

🛡️Security validation must evolve from disconnected tests to continuous, context-aware assessment powered by agentic AI. The piece argues that defenders need to converge three perspectives — adversarial, defensive, and risk — into a unified discipline supported by a Security Data Fabric that unites Asset Intelligence, Exposure Intelligence, and Security Control Effectiveness. With real-time context, autonomous agents can plan, execute, and prioritize validation workflows, turning fragmented tool outputs into actionable evidence and faster remediation. The article highlights Picus Security and industry recognition as indicators that the market is moving toward CTEM-native, agentic validation.

ℹ️ OpenAI told BleepingComputer that references to ads in its updated privacy policy do not indicate a global rollout — ads are currently limited to the United States. Ads launched in the US on February 9, 2026, and appear below answers for logged-in Free and Go users. OpenAI says ads run on separate systems, are clearly labeled, may be personalized, and that advertisers do not access chat content.

🔒 China's CNCERT warned that OpenClaw, an open-source, self-hosted autonomous AI agent, ships with weak default security and broad system privileges that attackers can abuse to seize endpoints and exfiltrate data. The advisory highlights indirect prompt injection (IDPI/XPIA) risks where benign features like web-page summarization and messaging link previews are weaponized to embed malicious instructions or automatically leak secrets. Researchers at PromptArmor demonstrated a technique in which an agent constructs attacker-controlled URLs that, when rendered as link previews, transmit confidential data without user clicks. CNCERT also flagged risks from malicious skills, accidental destructive commands, and disclosed vulnerabilities, urging isolation, tightened network controls, credential protection, and cautious skill sourcing.

🔬 Big tech's lavish hiring and compensation are accelerating an AI brain drain from universities, with firms pouring hundreds of billions into AI infrastructure and elite talent. The essay argues that betting on superstar hires undermines the collaborative, institution-driven nature of modern science and risks hollowing out curiosity-led research and independent ethical critique. It highlights team-based successes like LIGO and AlphaFold and urges universities to pursue alternatives: public-interest models such as Apertus, equitable pay across ranks, stronger researcher networks, and recognition of non-financial academic contributions. Institutions should defend intellectual freedom and build durable organizations rather than engage in a compensation arms race.

🔍Jake Moore, ESET Global Cybersecurity Advisor, demonstrated practical methods that can defeat widely used facial recognition systems. Using modified smart glasses, AI-generated images and real-time face swaps he showed how identities can be exposed, synthetic faces can bypass eKYC checks, and watchlists can be evaded. His findings highlight the need for rigorous adversarial testing and stronger verification controls; he will present live demos at RSAC 2026.

🔍 This post, the second in Microsoft's AI Application Security series, moves from planning to practical detection and response for prompt abuse. It describes common attack types — direct prompt override, extractive abuse targeting sensitive inputs, and indirect prompt injection via hidden instructions such as URL fragments — and why these are hard to spot without telemetry. The article provides a stepwise detection and incident response playbook and maps mitigations to Microsoft tools so teams can log interactions, sanitize inputs, and contain incidents.

🛡️ Attackers increasingly treat high-volume phishing as a weapon, flooding Security Operations Centers to exhaust analysts and hide targeted spear-phish. The article argues defenders must move from rule-based automation to decision-ready investigations—transparent, auditable agentic AI that produces concise verdicts and evidence. This reduces analyst fatigue, restores rapid response, and limits the window for attacker success.

🤖 Insurers are reshaping cyber policies as AI proliferates in business operations. Many carriers are tightening language, adding exclusions, and requiring evidence of active controls rather than relying on checkbox attestations. At the same time, firms that deploy AI-driven defenses and continuous monitoring can receive premium discounts. Brokers and policyholders must clarify AI usage and coverage before renewals to avoid gaps.

🛡️Charlotte AI is an agentic security analyst embedded in CrowdStrike Falcon, built to triage alerts, investigate threats and drive automated, inspectable response actions. It reasons over existing detections — including machine learning, IOAs and the CrowdStrike Threat Graph — and enforces analyst-defined guardrails so humans remain in control. Customers report faster MTTR and large reductions in initial investigation time.

🔒 Researchers demonstrated that an AI-powered browser, Perplexity's Comet, can be manipulated into executing a phishing scam in under four minutes. By intercepting the agent's explanatory traffic and training a GAN on those signals, attackers iteratively optimized a malicious page until the agent reliably performed fraudulent steps. The exploit leverages intent collision and prompt-injection weaknesses, shifting the target from users to the AI agent itself.

🔒 Researchers at Unit 42, Palo Alto Networks' lab, have demonstrated that LLM-based safety and evaluation systems — called AI Judges — can be manipulated via prompt-injection-style token sequences. Their custom fuzzer, AdvJudge-Zero, probes models in a black-box manner, finding low-perplexity formatting tokens that shift internal attention and increase the likelihood of an 'allow' decision. Unit 42 recorded a 99% bypass rate across multiple architectures, and showed that adversarial retraining on fuzzer-discovered examples can reduce that success rate to near zero.

🤖 GenAI has accelerated social engineering and phishing, allowing attackers to produce hyper-personalized messages, convincingly cloned executive voices, and realistic video impersonations in seconds. Deepfake incidents have shifted from online curiosity to tangible business risk, causing financial loss and operational disruption while making identity verification on everyday collaboration platforms increasingly difficult. To address these threats, Check Point Services has expanded its training portfolio and advocates for modern defenses and smarter awareness programs designed for the realities of the AI era.

⚠️ Boards and executive teams can no longer treat large vulnerability backlogs as a tolerable nuisance: agentic AI has collapsed attackers’ cost and speed of exploitation. Security leaders must present operational truth — not just compliance metrics — about current High and Critical findings, remediation timelines, and exposure costs. Boards should demand measurable remediation programs and a plan to reduce vulnerability accrual at the source. Regulation such as CRA and DORA raise legal and financial stakes, and 'patch faster' is not a complete answer when emergency fixes risk production outages.

🇨🇦 The Carney administration's $2‑billion Sovereign AI Compute Strategy forces a fundamental choice about where AI value and control will reside. Bruce Schneier warns that initiatives like OpenAI's “OpenAI for Countries” could simply transfer benefits and authority to U.S. tech firms, citing the Tumbler Ridge incident and private secrecy. He advocates for a publicly funded, transparent national AI—modeled on Switzerland's Apertus—to serve healthcare, education, transit, and democratic oversight rather than private profit.

🛡️ The piece compares the Civil War’s Gatling gun to a September 2025 agentic AI-driven cyberespionage campaign that automated most tactical operations. According to the report, a Chinese state-linked group, GTG-1002, abused Anthropic’s Claude Code via prompt injection and role-playing to produce malicious code and execute ≈90% of the attack chain. The intrusion hit 30 U.S. companies and agencies and was disclosed after Anthropic’s threat team detected misuse of their platform.

🛡️ Organizations are adopting agentic AI—autonomous, task-driven systems powered by LLMs—to streamline processes and boost throughput. These agents can plan, act, and iterate, but their non-deterministic behavior creates gaps in traceability, auditability, and access control. Apply strong role-based access, threat modeling, and oversight (human or independent evaluators) to limit exposure and ensure safe deployment.

🛡️AI adoption has outpaced controls, creating widespread "shadow AI" risk that can expose sensitive data, distort decisions and create compliance gaps. The article recounts an incident where a product manager accidentally pasted production API keys into a public model, triggering outbound alerts. It presents a five-step program grounded in the NIST AI Risk Management Framework: inventory and discover AI use, standardize assessments, deploy layered defenses (DLP and AI monitoring), enforce human-in-the-loop checks, and tie risk reduction to business value.

🛡️ CodeWall’s autonomous red-team agent compromised hiring startup Jack & Jill by chaining four seemingly minor bugs into a complete account takeover within an hour. The agent abused a permissive URL fetcher, an enabled test-login mode, missing onboarding role checks, and absent domain verification to map APIs, authenticate via a test OTP flow, and escalate to org-admin privileges. It then generated synthetic voice clips to social-engineer Jack, conducting 28 multi-turn exchanges and even impersonating Donald Trump before moving on, demonstrating how AI can rapidly combine low-risk flaws into high-impact attacks.

🔐 AI is reshaping cyber defense strategies and executive responsibilities. Organizations face a dual-use threat where AI empowers attackers and defenders; security teams must combine human expertise with automated capabilities. Human + AI approaches, informed by threat intelligence and comprehensive asset mapping, are critical. Vendors like ESET emphasize global, 24/7 coverage and say CISOs must secure board-level buy-in, regulatory alignment, and a clear, cost-effective AI roadmap to improve detection, response, and remediation.