Tag: training
-
Single prompt breaks AI safety in 15 major language models
Fundamental changes to safety mechanisms: The research went beyond measuring attack success rates to examine how the technique alters models’ internal safety mechanisms. When Microsoft tested Gemma3-12B-It on 100 diverse prompts, asking the model to rate their harmfulness on a 0-9 scale, the unaligned version systematically assigned lower scores, with mean ratings dropping from 7.97…
-
How to govern agentic AI so as not to lose control
assisting and start acting. We will witness a qualitative leap towards agent-based or agentive AI, capable of making autonomous decisions, managing complex workflows, and executing end-to-end tasks without constant intervention. However, this autonomy carries with it a serious warning for businesses: the ability to operate alone exponentially multiplies the impact of any error or security…
-
Leaked technical documents show China rehearsing cyberattacks on neighbors’ critical infrastructure
Internal files describe a training platform as part of a large integrated system designed to allow attackers to practice hacking replicas of “the real network environments” of China’s “main operational opponents in the South China Sea and Indochina directions.” First seen on therecord.media Jump to article: therecord.media/leaked-china-documents-show-testing-cyber-neighbors
-
Never settle: How CISOs can go beyond compliance standards to better protect their organizations
Tags: ai, awareness, breach, ciso, compliance, computing, control, cybersecurity, finance, risk, risk-assessment, risk-management, software, strategy, threat, training, vulnerabilityThe new North Star for CISOs: Accounting for emerging risk: We’ve established that it’s no longer good enough to overfit into a compliance standard, but you can still use compliance to your advantage.Most compliance programs mandate an information security risk assessment and, at a larger company, you may already have a dedicated enterprise risk management…
-
Gartner-Prognose: Die sechs wichtigsten Cybersicherheits-Trends für 2026
Tags: ai, awareness, business, compliance, computing, cyberattack, cybersecurity, cyersecurity, framework, gartner, governance, resilience, risk, soc, tool, trainingLesen Sie, mit welchen Cybersecurity-Trends sich Unternehmen in diesem Jahr beschäftigen sollten.Auch im Jahr 2026 bleibt die Cybersicherheitslage angespannt. Doch was sind die wichtigsten Themen, Risiken und Chancen, mit denen sich Security-Entscheider aktuell befassen sollten?Das Marktforschungsunternehmen Gartner hat dazu folgende sechs Trends ermittelt: KI-Agenten werden zunehmend von Mitarbeitern und Entwicklern genutzt, wodurch neue Angriffsflächen entstehen.…
-
KI als AWS-Angriffsturbo
Kriminelle Hacker haben ihre Angriffe auf AWS-Umgebungen mit KI beschleunigt.Forscher des Sicherheitsanbieters Sysdig haben einen Angriff aufgedeckt, bei dem kriminelle Angreifer eine AWS-Umgebung in weniger als acht Minuten vollständig kompromittieren konnten. Laut den Threat-Spezialisten nutzten die Bedrohungsakteure dabei eine Cloud-Fehlkonfiguration mit der Hilfe von Large Language Models (LLMs) aus, um den gesamten Angriffs-Lebenszyklus zu komprimieren…
-
The blind spot every CISO must see: Loyalty
Tags: access, ai, ciso, corporate, data, espionage, exploit, finance, framework, gartner, government, intelligence, jobs, malicious, monitoring, risk, strategy, tool, training, vulnerability, zero-trustHow the misread appears in practice: Recent examples illustrate the point. In the US federal sphere, abrupt terminations under workforce reduction initiatives have left former employees with lingering access to sensitive systems, amplifying the potential for data exposure or retaliation. Corporate cases show a similar dynamic: engineers or executives who have spent years building institutional…
-
Incognito Market admin sentenced to 30 years for running $105 million dark web drug empire
He promised “the best security there is” to hundreds of thousands of drug buyers, while quietly making the kind of mistake that guaranteed a 30-year sentence. And maybe training police on cryptocurrency while running a running a vast Tor-hidden drug bazaar wasn’t such a good idea. First seen on bitdefender.com Jump to article: www.bitdefender.com/en-us/blog/hotforsecurity/incognito-market-admin-sentenced-30-years-105-million-dark-web-drug-empire
-
Palo Alto Networks Revamps NextWave Program: 5 Biggest Updates
Palo Alto Networks is unveiling a major set of updates to its NextWave Partner Program including expanded incentives and new AI-powered training, along with increased expectations for top partners around driving platform adoption, Channel Chief Michael Khoury tells CRN. First seen on crn.com Jump to article: www.crn.com/news/security/2026/palo-alto-networks-revamps-nextwave-program-5-biggest-updates
-
When Documents Become the Attack Vector: Inside APT28’s Latest Microsoft Office Exploit
Email attachments remain one of the most trusted entry points into enterprise environments. Despite years of awareness training and secure email gateways, attackers continue to rely on documents because they blend seamlessly into everyday workflows. New reporting from The Hacker News details how APT28, a Russia-linked threat actor, is actively exploiting a newly disclosed Microsoft…
-
From credentials to cloud admin in 8 minutes: AI supercharges AWS attack chain
Tags: access, ai, attack, ciso, cloud, credentials, detection, framework, group, iam, least-privilege, LLM, monitoring, trainingLateral movement, LLMjacking, and GPU abuse: Once administrative access was obtained, the attacker moved laterally across 19 distinct AWS principals, assuming multiple roles and creating new users to spread activity across identities. This approach enabled persistence and complicated detection, the researchers noted.The attackers then shifted focus to Amazon Bedrock, enumerating available models and confirming that…
-
New phishing attack leverages PDFs and Dropbox
Masquerading as a safe document format: But after so many warnings about this over time, why are people still so trusting of PDFs and Dropbox?”Because, historically, they’ve actually been trained to be,” said Avakian. PDFs are routinely used in the business world and have been positioned as a safe, read-only document format for invoices, contracts,…
-
What’s New in Tenable Cloud Security: Multi-cloud Risk Analysis, Attack Surface Assessments, Improved IAM Security and More
Tags: ai, attack, cloud, compliance, data, data-breach, endpoint, gartner, google, governance, iam, identity, infrastructure, Internet, least-privilege, microsoft, mitigation, network, radius, risk, risk-analysis, service, supply-chain, switch, tool, training, vulnerabilityTenable Cloud Security continues to expand the technical depth of our Tenable One exposure management platform. Our latest enhancements include unified multi-cloud exploration, high-fidelity network validation, and expanded entitlement visibility across infrastructure and identity providers. Key takeaways Graph-based multi-cloud exploration: We’ve leveraged our unified data model to provide deep visibility across all cloud environments. You…
-
Das nächste große Security-Schlachtfeld
Tags: ai, chatgpt, computer, computing, cyber, cybersecurity, cyersecurity, encryption, framework, governance, Hardware, resilience, training, usaWenn Quantum Computing und KI in der Praxis zusammenkommen, bricht ein neues Zeitalter an auch und vor allem in Sachen Cybersecurity.In den letzten Jahren hat künstliche Intelligenz (KI) ihre Tentakel über die globale Technologielandschaft ausgebreitet. Das verdeutlicht unter anderem auch der zunehmende Einsatz von Automatisierung und autonomen Technologien in diversen Branchen und Sektoren. Und während…
-
Startup Amutable plotting Linux security overhaul to counter hacking threats
Tags: attack, backdoor, ceo, cloud, computer, computing, container, cve, cybercrime, data, exploit, fortinet, hacking, infrastructure, kubernetes, linux, microsoft, open-source, skills, software, startup, supply-chain, technology, threat, tool, training, vpn, vulnerabilitysystemd, he has alongside him two other ex-Microsoft employees, Chris Kühl as CEO, and Christian Brauner as CTO.A clue to Amutable’s plans lies in the announcement’s emphasis on some of its founders’ backgrounds in Kubernetes, runc, LXC, Incus, and containerd, all connected in different ways to the Linux container stack. Computing is full of security…
-
Human risk management: CISOs’ solution to the security awareness training paradox
Tags: access, ai, awareness, ciso, compliance, cyber, cybersecurity, data, email, identity, intelligence, malicious, mitigation, risk, risk-management, strategy, tool, trainingWhat is human risk management?: HRM is defined as a cybersecurity strategy that identifies, measures, and reduces the risks caused by human behavior. Simply stated, security awareness training is about what employees know; HRM is about what they do (i.e., their actual cybersecurity behavior).To be more specific, HRM integrates into email security tools, web gateways,…
-
Human risk management: CISOs’ solution to the security awareness training paradox
Tags: access, ai, awareness, ciso, compliance, cyber, cybersecurity, data, email, identity, intelligence, malicious, mitigation, risk, risk-management, strategy, tool, trainingWhat is human risk management?: HRM is defined as a cybersecurity strategy that identifies, measures, and reduces the risks caused by human behavior. Simply stated, security awareness training is about what employees know; HRM is about what they do (i.e., their actual cybersecurity behavior).To be more specific, HRM integrates into email security tools, web gateways,…
-
NDSS 2025 Reinforcement Unlearning
Session 10D: Machine Unlearning Authors, Creators & Presenters: Dayong Ye (University of Technology Sydney), Tianqing Zhu (City University of Macau), Congcong Zhu (City University of Macau), Derui Wang (CSIRO’s Data61), Kun Gao (University of Technology Sydney), Zewei Shi (CSIRO’s Data61), Sheng Shen (Torrens University Australia), Wanlei Zhou (City University of Macau), Minhui Xue (CSIRO’s Data61)…
-
Roughly half of employees are using unsanctioned AI tools, and enterprise leaders are major culprits
51% have connected AI tools to work systems or apps without the approval or knowledge of IT;63% believe it’s acceptable to use AI when there is no corporate-approved option or IT oversight;60% say speed is worth the security risk;21% think employers will simply “turn a blind eye” as long as they’re getting their work done.And…
-
NDSS 2025 TrajDeleter: Enabling Trajectory Forgetting In Offline Reinforcement Learning Agents
Session 10D: Machine Unlearning Authors, Creators & Presenters: hen Gong (University of Vriginia), Kecen Li (Chinese Academy of Sciences), Jin Yao (University of Virginia), Tianhao Wang (University of Virginia) PAPER TrajDeleter: Enabling Trajectory Forgetting in Offline Reinforcement Learning Agents Reinforcement learning (RL) trains an agent from experiences interacting with the environment. In scenarios where online…
-
THRM 2026: Changing Security Training and Human Behavior
Tags: trainingMost employees don’t ignore security training because they don’t care. They ignore it because it doesn’t feel real. A slide deck once a year. A short quiz at the end. A reminder to “stay vigilant.” These formats are familiar, easy to roll out, and easy to measure. They are also poorly suited for the moments……
-
Crooks are hijacking and reselling AI infrastructure: Report
Tags: access, ai, api, attack, authentication, business, cloud, communications, control, credentials, cybersecurity, data, data-breach, endpoint, exploit, firewall, group, infosec, infrastructure, intelligence, Internet, LLM, malicious, marketplace, risk, service, skills, technology, theft, threat, training, vulnerabilityexposed endpoints on default ports of common LLM inference services;unauthenticated API access without proper access controls;development/staging environments with public IP addresses;MCP servers connecting LLMs to file systems, databases and internal APIs.Common misconfigurations leveraged by these threat actors include:Ollama running on port 11434 without authentication;OpenAI-compatible APIs on port 8000 exposed to the internet;MCP servers accessible without…
-
I’m a tech-savvy zillennial who knows how to safeguard against hacking. Scammers still managed to get me | Caitlin Cassidy
Had I received any suspicious text messages claiming to be from my bank, the fraud team asked. Had I clicked on the links? My stomach dropped<ul><li>Get our <a href=”https://www.theguardian.com/email-newsletters?CMP=cvau_sfl”>breaking news email, <a href=”https://app.adjust.com/w4u7jx3″>free app or <a href=”https://www.theguardian.com/australia-news/series/full-story?CMP=cvau_sfl”>daily news podcast</li></ul>The scariest part about getting scammed was not realising it was happening in the first place.Perhaps naively,…
-
Skills CISOs need to master in 2026
Tags: access, ai, business, ciso, cloud, compliance, credentials, cyber, cybersecurity, data, endpoint, finance, firewall, group, Hardware, identity, infrastructure, intelligence, jobs, resilience, risk, risk-management, skills, strategy, threat, tool, trainingTop technical skills: In addition to strong knowledge of AI systems, today’s CISOs need a solid foundation in the technologies that define modern enterprise environments. The (ISC)² CISSP is still widely regarded as the gold standard for broad expertise in security architecture, risk management, and governance. “Regulators will expect this, and it still appears in…
-
Overcoming AI fatigue
Tags: access, ai, awareness, business, ciso, cloud, control, data, finance, governance, incident response, jobs, metric, monitoring, privacy, risk, strategy, supply-chain, technology, tool, training, zero-trustbefore it becomes fully entrenched in every corner of the business. It’s a rare opportunity, one we shouldn’t waste. A big part of the confusion comes from the word “AI” itself. We use the same label to talk about a chatbot drafting marketing copy and autonomous agents that generate and implement incident response playbooks. Technically,…