“`html
Cybersecurity investigators have uncovered what they claim to be the most primitive instance recognized to date of malicious software that incorporates Large Language Model (LLM) functionalities.
The malware has been labeled MalTerminal by the SentinelOne SentinelLABS research group. The results were showcased at the LABScon 2025 security symposium.
In a document analyzing the nefarious application of LLMs, the cybersecurity firm
“`