darkreading

A new around of vulnerabilities in the popular AI automation platform could let attackers hijack servers and steal credentials, allowing full takeover.

If an attacker splits a malicious prompt into discrete chunks, some large language models (LLMs) will get lost in the details and miss the true intent.

In the latest edition of "Reporters' Notebook," a trio of journalists urge the cybersecurity industry to prioritize patching vulnerabilities, preparing for quantum threats, and refining AI applications,

Ransomware defense requires focusing on business resilience. This means patching issues promptly, improving user education, and deploying multifactor authentication.

Russian and Chinese nation-state attackers are exploiting a months-old WinRAR vulnerability, despite a patch that came out last July.

To stop the ongoing attacks, the cybersecurity vendor took the drastic step of temporarily disabling FortiCloud single sign-on (SSO) authentication for all devices.

The retail sector must adapt as consumers become more cybersecurity-conscious. Increased attack transparency is a good place to start.

In two separate campaigns, attackers used the JScript C2 framework to target Chinese gambling websites and Asian government entities with new backdoors.

The region is up against tactics like data-leak extortion, credential-stealing campaigns, edge-device exploitation, and attackers leveraging AI.

AI "model collapse," where LLMs over time train on more and more AI-generated data and become degraded as a result, can introduce inaccuracies, promulgate malicious activity, and impact PII protections.

A new ransomware strain that entered the scene last year has poorly designed code and uses Hebrew language that might be a false flag.

While telnet is considered obsolete, the network protocol is still used by hundreds of thousands of legacy systems and IoT devices for remote access.

To exploit the vulnerability, an attacker would need either system access or be able to convince a user to open a malicious Office file.

The malware-as-a-service kit enables malicious extensions to overlay pages on real websites without changing the visible URL, signaling a fresh challenge for enterprise security.

The sportswear brand is investigating an alleged breach of its network that exposed some 188,347 files of highly sensitive corporate data.

The sportswear brand is investigating an alleged breach of its network that exposed some 188,347 files of highly sensitive corporate data.

How MITRE has mismanaged the world's vulnerability database for decades and wasted millions along the way.

Artistic initiatives turn cybersecurity into immersive exhibits at the Museum of Malware Art, transforming digital threats into thought-provoking experiences.

Researchers attributed the failed attempt to the infamous Russian APT Sandworm, which is notorious for wiper attacks on critical infrastructure organizations.

The North Korean threat group is using a new PowerShell backdoor to compromise development environments and target cryptocurrency holdings, according to researchers.