Tag: encryption
-
LLM side-channel attack could allow snoops to guess what you’re talking about
Encryption protects content, not context First seen on theregister.com Jump to article: www.theregister.com/2025/11/11/llm_sidechannel_attack_microsoft_researcher/
-
New Attacks Against Secure Enclaves
Encryption can protect data at rest and data in transit, but does nothing for data in use. What we have are secure enclaves. I’ve written about this before: Almost all cloud services have to perform some computation on our data. Even the simplest storage provider has code to copy bytes from an internal storage system…
-
New Attacks Against Secure Enclaves
Encryption can protect data at rest and data in transit, but does nothing for data in use. What we have are secure enclaves. I’ve written about this before: Almost all cloud services have to perform some computation on our data. Even the simplest storage provider has code to copy bytes from an internal storage system…
-
New Attacks Against Secure Enclaves
Encryption can protect data at rest and data in transit, but does nothing for data in use. What we have are secure enclaves. I’ve written about this before: Almost all cloud services have to perform some computation on our data. Even the simplest storage provider has code to copy bytes from an internal storage system…
-
New Whisper-Based Attack Reveals User Prompts Hidden Inside Encrypted AI Traffic
Microsoft researchers have unveiled a sophisticated side-channel attack targeting remote language models that could allow adversaries to infer conversation topics from encrypted network traffic. Despite end-to-end encryption via Transport Layer Security (TLS), the attack exploits patterns in packet sizes and timing to classify the subject matter of user prompts sent to AI chatbots. The research…
-
New Whisper-Based Attack Reveals User Prompts Hidden Inside Encrypted AI Traffic
Microsoft researchers have unveiled a sophisticated side-channel attack targeting remote language models that could allow adversaries to infer conversation topics from encrypted network traffic. Despite end-to-end encryption via Transport Layer Security (TLS), the attack exploits patterns in packet sizes and timing to classify the subject matter of user prompts sent to AI chatbots. The research…
-
AI chat privacy at risk: Microsoft details Whisper Leak side-channel attack
Microsoft uncovered Whisper Leak, a side-channel attack that lets network snoopers infer AI chat topics despite encryption, risking user privacy. Microsoft revealed a new side-channel attack called Whisper Leak, which lets attackers who can monitor network traffic infer what users discuss with remote language models, even when the data is encrypted. The company warned that…
-
Microsoft Uncovers ‘Whisper Leak’ Attack That Identifies AI Chat Topics in Encrypted Traffic
Microsoft has disclosed details of a novel side-channel attack targeting remote language models that could enable a passive adversary with capabilities to observe network traffic to glean details about model”¯conversation”¯topics despite encryption protections under certain circumstances.This leakage of data exchanged between humans and”¯streaming-mode language models could pose serious risks to First seen on thehackernews.com Jump…
-
Gesundheitswesen: Erpressungen auf Höchststand, Stress bei den Teams
Schwachstellen und Kapazitätsprobleme sind die Hauptursachen für Ransomware-Angriffe im Gesundheitswesen, wobei die Verschlüsselung von Daten auf ein Fünfjahrestief gesunken ist. Gleichzeitig sind die Lösegeldforderungen und -zahlungen zurückgegangen, während die Angreifer vermehrt auf reine Erpressungsangriffe setzen. Die Belastung der IT- und Cybersicherheitsteams ist gestiegen, was sich in erhöhtem Druck, Angst und Stress äußert. In der aktuellen……
-
Gesundheitswesen: Erpressungen auf Höchststand, Stress bei den Teams
Schwachstellen und Kapazitätsprobleme sind die Hauptursachen für Ransomware-Angriffe im Gesundheitswesen, wobei die Verschlüsselung von Daten auf ein Fünfjahrestief gesunken ist. Gleichzeitig sind die Lösegeldforderungen und -zahlungen zurückgegangen, während die Angreifer vermehrt auf reine Erpressungsangriffe setzen. Die Belastung der IT- und Cybersicherheitsteams ist gestiegen, was sich in erhöhtem Druck, Angst und Stress äußert. In der aktuellen……
-
How TeamViewer builds enterprise trust through security-first design
Tags: encryptionWhat to do when even your espresso machine needs end-to-end encryption First seen on theregister.com Jump to article: www.theregister.com/2025/11/07/teamviewer-security-first-design/
-
Digital health can’t scale if cybersecurity falls behind
Tags: access, ai, attack, breach, cloud, compliance, control, cyber, cyberattack, cybersecurity, data, detection, encryption, endpoint, exploit, framework, GDPR, governance, government, healthcare, HIPAA, identity, infection, intelligence, malicious, network, nist, phishing, privacy, ransomware, regulation, resilience, risk, risk-management, strategy, technology, threat, training, virus, vulnerability, zero-trustThe unique vulnerabilities of AI systems: Traditional security frameworks are not enough for AI. Attacks on algorithms take subtler forms. I often explain to my clients that when you corrupt data, you corrupt intelligence. Data poisoning occurs when malicious data is inserted into the training process, teaching the AI to make wrong decisions later. Imagine…
-
Digital health can’t scale if cybersecurity falls behind
Tags: access, ai, attack, breach, cloud, compliance, control, cyber, cyberattack, cybersecurity, data, detection, encryption, endpoint, exploit, framework, GDPR, governance, government, healthcare, HIPAA, identity, infection, intelligence, malicious, network, nist, phishing, privacy, ransomware, regulation, resilience, risk, risk-management, strategy, technology, threat, training, virus, vulnerability, zero-trustThe unique vulnerabilities of AI systems: Traditional security frameworks are not enough for AI. Attacks on algorithms take subtler forms. I often explain to my clients that when you corrupt data, you corrupt intelligence. Data poisoning occurs when malicious data is inserted into the training process, teaching the AI to make wrong decisions later. Imagine…
-
Digital health can’t scale if cybersecurity falls behind
Tags: access, ai, attack, breach, cloud, compliance, control, cyber, cyberattack, cybersecurity, data, detection, encryption, endpoint, exploit, framework, GDPR, governance, government, healthcare, HIPAA, identity, infection, intelligence, malicious, network, nist, phishing, privacy, ransomware, regulation, resilience, risk, risk-management, strategy, technology, threat, training, virus, vulnerability, zero-trustThe unique vulnerabilities of AI systems: Traditional security frameworks are not enough for AI. Attacks on algorithms take subtler forms. I often explain to my clients that when you corrupt data, you corrupt intelligence. Data poisoning occurs when malicious data is inserted into the training process, teaching the AI to make wrong decisions later. Imagine…
-
Why can’t enterprises get a handle on the cloud misconfiguration problem?
Tags: access, ai, authentication, awareness, breach, business, cloud, communications, computing, control, cybersecurity, data, data-breach, encryption, governance, hacker, infrastructure, least-privilege, mfa, monitoring, network, risk, saas, service, technology, tool, training, usa, zero-trustStop. Reassess. Reconfigure: Last year, according to Ayan Roy, EY Americas cybersecurity competency leader, the highest number of breaches were caused by shared cloud repositories. “That’s where we saw the maximum amount of data exfiltration,” he says. “A lot was from shared cloud stores and SaaS applications.” That’s despite the fact that the clients have…
-
Why can’t enterprises get a handle on the cloud misconfiguration problem?
Tags: access, ai, authentication, awareness, breach, business, cloud, communications, computing, control, cybersecurity, data, data-breach, encryption, governance, hacker, infrastructure, least-privilege, mfa, monitoring, network, risk, saas, service, technology, tool, training, usa, zero-trustStop. Reassess. Reconfigure: Last year, according to Ayan Roy, EY Americas cybersecurity competency leader, the highest number of breaches were caused by shared cloud repositories. “That’s where we saw the maximum amount of data exfiltration,” he says. “A lot was from shared cloud stores and SaaS applications.” That’s despite the fact that the clients have…
-
Why can’t enterprises get a handle on the cloud misconfiguration problem?
Tags: access, ai, authentication, awareness, breach, business, cloud, communications, computing, control, cybersecurity, data, data-breach, encryption, governance, hacker, infrastructure, least-privilege, mfa, monitoring, network, risk, saas, service, technology, tool, training, usa, zero-trustStop. Reassess. Reconfigure: Last year, according to Ayan Roy, EY Americas cybersecurity competency leader, the highest number of breaches were caused by shared cloud repositories. “That’s where we saw the maximum amount of data exfiltration,” he says. “A lot was from shared cloud stores and SaaS applications.” That’s despite the fact that the clients have…
-
Why can’t enterprises get a handle on the cloud misconfiguration problem?
Tags: access, ai, authentication, awareness, breach, business, cloud, communications, computing, control, cybersecurity, data, data-breach, encryption, governance, hacker, infrastructure, least-privilege, mfa, monitoring, network, risk, saas, service, technology, tool, training, usa, zero-trustStop. Reassess. Reconfigure: Last year, according to Ayan Roy, EY Americas cybersecurity competency leader, the highest number of breaches were caused by shared cloud repositories. “That’s where we saw the maximum amount of data exfiltration,” he says. “A lot was from shared cloud stores and SaaS applications.” That’s despite the fact that the clients have…
-
DragonForce Cartel Surfaces from Leaked Conti v3 Ransomware Source Code
Tags: cyber, data-breach, encryption, flaw, group, malware, ransomware, software, threat, vulnerabilityAcronis Threat Research Unit has analyzed recent activity linked to the DragonForce ransomware group and identified a new malware variant in the wild. The latest sample uses vulnerable drivers such as truesight.sys and rentdrv2.sys to disable security software, terminate protected processes and correct encryption flaws previously associated with Akira ransomware. The updated encryption scheme addresses…
-
Check Point knackt XLoader-Malware mit Generativer KI
Bei der Untersuchung von XLoader stießen die Forscher unter anderem auf eine mehrschichtige RC4-Verschlüsselung, versteckte Windows-API-Aufrufe und neue Mechanismen zur Umgehung von Sandbox-Umgebungen. First seen on infopoint-security.de Jump to article: www.infopoint-security.de/check-point-knackt-xloader-malware-mit-generativer-ki/a42609/
-
XLoader Malware Analyzed Using ChatGPT’s AI, Breaks RC4 Encryption Layers in Hours
Cybersecurity researchers have successfully demonstrated how artificial intelligence can dramatically accelerate malware analysis, decrypting complex XLoader samples in a fraction of the time previously required. XLoader, a sophisticated malware loader with information-stealing capabilities dating back to 2020, has long been considered one of the most challenging malware families to analyze. The malware combines multiple layers…
-
XLoader Malware Analyzed Using ChatGPT’s AI, Breaks RC4 Encryption Layers in Hours
Cybersecurity researchers have successfully demonstrated how artificial intelligence can dramatically accelerate malware analysis, decrypting complex XLoader samples in a fraction of the time previously required. XLoader, a sophisticated malware loader with information-stealing capabilities dating back to 2020, has long been considered one of the most challenging malware families to analyze. The malware combines multiple layers…
-
New BOF Tool Bypasses Microsoft Teams Cookie Encryption to Steal User Chats
Cybersecurity researchers at Tier Zero Security have released a specialised Beacon Object File (BOF) tool that exploits a critical weakness in Microsoft Teams cookie encryption, enabling attackers to steal user chat messages and other sensitive communications. The vulnerability stems from how Microsoft Teams handles cookie encryption compared to modern Chromium-based browsers. While contemporary browsers like…
-
Preventing DNS filtering bypass by Encrypted DNS (DoT, DoH, DoQ)
DNS over HTTPS (DoH) and other encrypted DNS protocols like DNS over TLS (DoT) & DNS over QUIC (DoQ) enhances user privacy and security by encrypting DNS queries in transit, shielding them from eavesdropping, tampering, and censorship on untrusted networks. This prevents ISPs and local attackers from logging or manipulating domain resolutions, fostering a more…
-
Cyber agencies produce ‘long overdue’ best practices for securing Microsoft Exchange Server
The guidance: The guidance states admins should treat on-prem Exchange servers as being “under imminent threat,” and itemizes key practices for admins:First, it notes, “the most effective defense against exploitation is ensuring all Exchange servers are running the latest version and Cumulative Update (CU)”;It points out that Microsoft Exchange Server Subscription Edition (SE) is the…
-
Cyber agencies produce ‘long overdue’ best practices for securing Microsoft Exchange Server
The guidance: The guidance states admins should treat on-prem Exchange servers as being “under imminent threat,” and itemizes key practices for admins:First, it notes, “the most effective defense against exploitation is ensuring all Exchange servers are running the latest version and Cumulative Update (CU)”;It points out that Microsoft Exchange Server Subscription Edition (SE) is the…