All news with #research tag

🔍 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.

🔒 A recent PPI survey of 50 banks and 53 insurers in Germany reports a sixfold rise in cyberattacks compared with 2021. Sixty-four percent of respondents now view cyberattacks as the sector's top challenge, ahead of digitization, credit quality and regulation. Firms cite low employee awareness and difficulty with real-time detection; malware installation and IT disruption are the most frequent attack types.

🛡️ Roughly 70% of senior security leaders say internal conflicts during a cyber crisis cause more disruption than the attack itself, according to the Cytactic 2025 State of Cyber Incident Response Management (CIRM) Report. The survey of 480 US cybersecurity executives highlights blurred authority, poor communication, and unrehearsed roles that delay response. Experts recommend demonstrating security's business value, reducing operational friction with passwordless controls, and aligning incentives with lines of business.

📊 Cloudflare explains why measuring the Internet is uniquely difficult and how rigorous methodology, ethics, and clear representation make findings reliable. An internal February 2022 Lviv traffic spike illustrates how context and complementary data can prevent misclassification of benign events as attacks. The post contrasts active and passive techniques and direct versus indirect measurement, outlines a lifecycle of curation, modeling, and validation, and stresses low-impact, ethical approaches. It concludes by inviting collaboration and continued exploration of passive measurement methods.

🔍 During a 2022 internship at Cloudflare, Ram Sundara Raman examined whether connection tampering by network middleboxes can be detected using only passive production data. He sampled one in 10,000 TCP connections and logged the first ten inbound packets, then developed 19 tampering signatures while confronting scale, noisy telemetry, and limited ground truth. The work exposed practical limits of passive observation and the care required to interpret packet-level signals, and its outputs are published on Cloudflare Radar.

📡 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.

🔒 Incogni's Social Media Privacy Ranking 2025, summarized by Kaspersky, evaluates 15 platforms across 18 criteria to compare messaging apps on privacy and data handling. Overall scores place Discord, Telegram and Snapchat near the top, but a subset of practical criteria ranks Telegram first, followed by Snapchat and Discord. The analysis highlights default settings, data collection by mobile apps, handling of government requests, and encryption differences, noting that only WhatsApp provides end-to-end encryption for all chats by default.

🔬 Researchers at the Max Planck Institute of Biochemistry and Google Cloud created a Proteomics Lab Agent using the Agent Development Kit and Gemini models to provide personalized, multimodal AI guidance for mass spectrometry experiments. The agent analyzes recorded steps to generate publication-ready protocols, detect procedural errors, and capture tacit expertise into a searchable knowledge base. Open-sourced on GitHub, it aims to reduce troubleshooting time and improve reproducibility across labs.

🔒 Cybersecurity researchers and rights groups warn the UN Convention against Cybercrime, which begins a ratification process in Hanoi this weekend, could criminalize legitimate research and expand intrusive surveillance powers. The Cybersecurity Tech Accord and organizations such as Human Rights Watch say the draft's vague scope, broad criminalization language, and expansive data-access provisions risk arbitrary abuse and could hamper incident response. Some analysts acknowledge improvements around intent-based language but stress that robust national safeguards and explicit protections for security research are still needed.

🛡️ Pwn2Own Ireland 2025 concluded in Cork with security researchers awarded $1,024,750 after demonstrating 73 zero-day vulnerabilities across eight product categories. Targets included printers, network-attached storage, messaging apps, smart home and surveillance devices, home networking gear, flagship phones (iPhone 16, Galaxy S25, Pixel 9) and wearables. The contest expanded the attack surface to include USB port exploitation on locked mobile handsets while retaining Bluetooth, Wi‑Fi and NFC vectors. Summoning Team topped the leaderboard with $187,500 and 22 Master of Pwn points.

🔐 A high-profile entry by a hacker known as ‘Eugene’ at Pwn2Own Ireland 2025 withdrew a claimed zero-click remote code execution exploit targeting WhatsApp, forfeiting the event’s $1 million top prize. Organizers Trend Micro ZDI say Team Z3 is sharing findings privately for coordinated disclosure to Meta, while WhatsApp reports no viable exploit was publicly demonstrated. The cancellation has fueled speculation about exploit readiness and underscores the role of responsible disclosure and rigorous triage before public demonstrations.

🔊 A recent study by researchers at the University of California, Irvine shows that very high-resolution optical sensors in some mice can detect minute desk vibrations produced by speech. The theoretical attack, labeled Mic-E-Mouse, requires mice with extremely high DPI (≈10,000+) and very high polling rates (≈4,000 Hz+) and malware to exfiltrate raw sensor frames. The raw signals are extremely noisy, but Wiener filtering and ML-based denoising allowed partial speech recovery under controlled lab conditions. Significant practical limitations — few qualifying models, controlled setups with speakers inches from the sensor, and steep drops in accuracy with common barriers — plus straightforward mitigations make the attack largely a proof of concept for now.

⚠️ Researchers at SquareX demonstrated an AI Sidebar Spoofing attack that can overlay a counterfeit assistant in OpenAI's Atlas and Perplexity's Comet browsers. A malicious extension injects JavaScript to render a fake sidebar identical to the real UI and intercepts all interactions, leaving users unaware. SquareX showcased scenarios including cryptocurrency phishing, OAuth-based Gmail/Drive hijacks, and delivery of reverse-shell installation commands. The team reported the findings to vendors but received no response by publication.

⚠️ The latest releases of Cursor and Windsurf IDEs embed outdated Chromium and V8 engines that contain at least 94 known, patched vulnerabilities. Ox Security researchers demonstrated a proof‑of‑concept exploiting CVE-2025-7656 (a Maglev JIT integer overflow) to crash Cursor, and warn that similar flaws could enable denial‑of‑service or arbitrary code execution in real attacks. Attack vectors include deeplinks, malicious extensions, poisoned README previews or documentation; the two IDEs together serve an estimated 1.8 million developers. Cursor dismissed the DoS finding as out of scope and Windsurf did not respond to inquiries.

🔒On the first day of Pwn2Own Ireland 2025, security researchers exploited 34 unique zero-day vulnerabilities and collected $522,500 in cash awards. Team DDOS (Bongeun Koo and Evangelos Daravigkas) chained eight flaws to compromise a QNAP Qhora-322 router via its WAN interface and access a QNAP TS-453E, earning $100,000 and moving into second place on the Master of Pwn leaderboard. The Summoning Team led day one with $102,500 and 11.5 points after multiple successful root exploits. The Zero Day Initiative (ZDI) organized the event and coordinates 90-day responsible disclosure with affected vendors.

🧠 In episode 73 of The AI Fix, hosts Graham Cluley and Mark Stockley explore a sweep of recent AI developments, from the rise of AI-generated content to high-profile figures relying on chatbots. They discuss research suggesting Google Gemini exhibits behaviours resembling pathological gambling and report on a Gemma-style model uncovering a potential cancer therapy pathway. The show also highlights legal and security concerns— including a lawyer criticised for repeated AI use, generals consulting chatbots, and techniques for poisoning LLMs with only a few malicious samples.

🔍 Cloudflare investigated a production soft lockup traced to the Linux BPF LPM trie, a core data structure for IP and IP+Port longest-prefix matching. Benchmarks on 96-core AMD EPYC hardware showed lookups remain relatively fast at modest sizes, but updates, deletes and especially freeing maps degrade severely at scale, causing multi-second CPU stalls and customer packet loss. The post refreshes trie basics, presents measured results (lookups, updates, deletes, free costs), and diagnoses kernel implementation limits — notably binary child pointers, absent level compression, and allocator-induced cache and dTLB pressure — then outlines plans to upstream benchmarks and refactor toward a level-compressed multibit trie to reduce traversal height, cache/TLB misses, and freeing overhead.

🔍 This article frames the shift from legacy SOCs to AI-SOC platforms, arguing leaders must evaluate impact, transparency, and integration rather than pursue AI for its own sake. It outlines four architectural dimensions—functional domain, implementation model, integration architecture, and deployment—and prescribes a phased adoption path with concrete vendor questions. The piece flags key risks including explainability gaps, data residency, vendor lock-in, model drift, and cost surprises, and highlights mitigation through governance, human-in-the-loop controls, and measurable POCs.

🔐 The arrival of cryptographically relevant quantum computers will create a "silent boom" where adversaries can capture encrypted traffic today and decrypt it later, making intrusions neither observed nor observable. This undermines traditional incident response and shifts responsibility to engineering teams, not a vendor checkbox. Organizations must pursue quantum readiness by engaging developers to inventory algorithms and data, assess internet-facing assets for PQC support, and build testing capability for new ciphers within their release cycles.

🛡️ Penetration testing remains a critical control, but the classic, one-size-fits-all approach can create hidden financial and operational burdens. Administrative overheads, complex scoping decisions and indirect remediation work all add time and cost while risking scope creep and disruption. The article recommends flexible, consumption-based models—such as PTaaS and Outpost24's CyberFlex—to improve coverage, transparency and ROI.