New enhancements to the Sophos AI Assistant

AI agents are rapidly becoming a core part of the enterprise, being embedded across enterprise workflows, operating with autonomy, and making decisions about which systems to access and how to use them. But as agents grow in power and autonomy, so do the risks and threats.  Recent studies show 80% of companies have already experienced…

The North Korean threat actor linked to the Contagious Interview campaign has been observed merging some of the functionality of two of its malware programs, indicating that the hacking group is actively refining its toolset. That’s according to new findings from Cisco Talos, which said recent campaigns undertaken by the hacking group have seen the…

Konni, a subset of the state-sponsored DPRK cyberespionage group, first exploits Google Find Hub, which ironically aims to protect lost Android devices, to remotely wipe devices.

Experts recommend enhanced AI literacy, training around the ethics of using AI, and verification protocols to maintain credibility in an increasingly AI-influenced courtroom.

Various generative artificial intelligence (GenAI) services have been found vulnerable to two types of jailbreak attacks that make it possible to produce illicit or dangerous content. The first of the two techniques, codenamed Inception, instructs an AI tool to imagine a fictitious scenario, which can then be adapted into a second scenario within the first…

A Chinese-speaking threat actor tracked as UAT-6382 has been linked to the exploitation of a now-patched remote-code-execution vulnerability in Trimble Cityworks to deliver Cobalt Strike and VShell. “UAT-6382 successfully exploited CVE-2025-0944, conducted reconnaissance, and rapidly deployed a variety of web shells and custom-made malware to maintain long-term access,” Cisco Talos researchers