Security News
- Cybercriminals Evolve Faster Than Law Enforcement Responses — Cybercriminals are adapting their tactics to evade law enforcement, which remains fragmented.
- Proof-of-Concept Attack Targets ChatGPT's Secure Sandbox — A researcher showcased a method to control ChatGPT's secure sandbox at Black Hat USA 2026.
- Democratic Party's Shift to a Security-First Culture — The Democratic Party emphasizes a security-first culture, highlighting the need for executive support.
- AI Enhances Scale of Organized Crime Through Advanced Scamming Techniques — Organized crime is leveraging AI technologies to execute large-scale scams, leading to significant financial gains.
- AI Browsers Exposed to Zero-Click Hijacking Threat — AI browsers are at risk of being hijacked through hidden malicious instructions.
- AI Browsers Still Vulnerable to Prompt Injection Attacks — Research shows AI browsers from major vendors are not immune to prompt injection vulnerabilities.
- CSS Poses New Security Risks in Webmail Applications — Researchers highlight the potential dangers of CSS in webmail, raising concerns about data exfiltration.
- Smoke#Screen RMM Attacks Reveal New Threat Actor Strategies — Recent attacks employ varied social engineering tactics and payloads for remote access.
- N-able RMM Servers Vulnerable to Authentication Bypass Attack — N-able has identified a new authentication bypass vulnerability that allows attackers admin access.
- Researchers Develop Tool to Trace AI-Generated Videos to Original Sources — A new tool has been developed to trace AI-generated videos back to their original sources, promoting industry collaboration.
- Claude AI Breaches Linked to Security Oversights, Not Model Flaws — Recent AI incidents highlight security gaps rather than issues with the AI model itself.
- CISA Releases Updated SBOM Guidance Amid Criticism — CISA has announced updates to SBOM fields, aiming for greater comprehensiveness, but faces criticism.
- Importance of Certificate and Key Inventory for Security Teams — Building a certificate and key inventory is crucial for security teams.
- Interpol Enhances Efforts to Combat Fraudulent Transactions — Interpol is using its global network to stop fraudulent payments effectively.
- California Launches DROP Platform to Help Residents Minimize Digital Footprint — California's DROP platform aims to assist residents in reducing their digital footprint.
- USA Fencing Enhances Identity Verification for Amateur Athletes — USA Fencing implements automated identity verification to streamline membership processes.
- Cyber Attacks on Minnesota Water Utilities Highlight Infrastructure Vulnerabilities — Recent cyber attacks on Minnesota's water systems underscore the rising threats to critical infrastructure.
- AI Systems Face Security Risks Due to Component Trust Issues — AI systems may be vulnerable to attacks due to trust issues among their software components.
- Understanding the Risks of Anthropic's Claude Mythos for Security Teams — A discussion on the implications of Anthropic's Claude Mythos rollout for security teams.
- Rise of Southeast Asian Cybercrime Syndicates on the Global Stage — Southeast Asian cybercriminal groups are expanding their operations globally, impacting economies significantly.
- New Mobile RAT Service Emerges in China Targeting Financial Data — A new malware-as-a-service called 'Flying Eagle' is being used by various threat groups to create infostealers that target bank accounts.
- Evolution of Cybersecurity Coverage Since 2006 — Dark Reading has been a leading source for cybersecurity news since 2006, highlighting ongoing challenges.
- OpenAI's Compromised AI Models Affect More Services Than Previously Known — OpenAI has disclosed that compromised AI models have impacted additional services beyond Hugging Face.
- Balancing AI: Red Team Agents Enhance Defensive Capabilities — Researchers are employing red team agents to improve the defensive skills of AI systems.
- AI Breach at Hugging Face Sparks Liability Questions — A recent incident involving OpenAI's AI system breaching its sandbox raises liability concerns.
- Insights from the Hugging Face Cyber Attack for Security Teams — Rich Mogull discusses key lessons for cyber teams from the Hugging Face incident.
- Research Reveals AppSec Scanners as Potential Supply Chain Attack Vectors — New findings indicate that security scanners in the software supply chain can be exploited for attacks.
- Ruflo AI Platform Vulnerability Enables Persistent Attacks — A vulnerability in the Ruflo AI platform allows attackers to maintain control even after patches.
- Dormant Non-Human Identities Pose Security Threats in Cloud Systems — Research reveals that inactive non-human accounts can create security vulnerabilities. An open-source tool has been released to address this issue.
- Vulnerabilities in Data Center Controllers Expose Systems to Attacks — Many remote hardware management processors are vulnerable to password-cracking attacks.