DARKReading

Despite adding alignment training, guardrails, and filters, large language models continue to give up secrets, make unfiltered statements, and provide dangerous information.

Prompt injection risks in GitLab's AI assistant could have allowed attackers to steal source code, or indirectly deliver developers malware, dirty links, and more.

A recent spear-phishing campaign against countries in South Asia aligns with broader political tensions in the region.

Gaps in laws, technology, and corporate accountability continue to put women's safety and privacy online at risk.

The FBI and partners have disrupted "the world's most popular malware," a sleek enterprise with thousands of moving parts, responsible for millions of cyberattacks in every part of the world.

Wiz researchers found an opportunistic threat actor has been targeting vulnerable edge devices, including Ivanti VPNs and Palo Alto firewalls.

The company expects it will continue to struggle with online disruptions until at least July, due to the attack.

Vixen Panda, Aquatic Panda — both Beijing-sponsored APTs and financially motivated criminal groups continued to pose the biggest threat to organizations in Central and South America last year, says CrowdStrike.

A number of major industrial organizations suffered ransomware attacks last quarter, such as PCB manufacturer Unimicron, appliance maker Presto, and more — a harbinger of a rapidly developing and diversifying threat landscape.

Coinbase asserts that this number is only a small fraction of the number of its verified users, though it's still offering a $20 million reward to catch the criminals.