Following the emergence of large language model-based (LLM) chatbots for the consumer market, organizations increasingly incorporate similar tools for business purposes.
Email security provider Ironscales has launched a GPT-powered chat assistant for self-service threat reporting on Microsoft Outlook during Infosecurity Europe.
The machine learning-powered application, called Themis Co-pilot, aims to reduce the number of user-reported false positives and automatically classify and remediate threats after users reported similar ones.
Themis Co-pilot is based on PhishLLM, a patent-pending custom large language model hosted and developed by Ironscales and will be enriched by Themis AI, the vendor’s AI-powered security analyst.
Launched in 2018, Themis AI collects data from millions of security events from users, devices, and threat intel signals in a continuous reinforcement learning from human feedback (RLHF) cycle, as well as human insights collected from every mailbox user and over 20,000 security analysts across the Ironscales network of global admins.
With Themis Co-pilot, however, Ironscales goes further by enabling anyone to use AI for security purposes, IRONSCALES’ CEO Eyal Benishti said in a public statement: “Ironscales has pioneered the use of AI in email security to detect and remediate sophisticated phishing attempts. With the introduction of Themis Co-pilot, we’re delivering the next innovation that will help end users, of any skill level, improve their ability to stop attacks without adding additional cost or complexity to the organization.”
The announcement comes when rates of advanced phishing attacks are at an all-time high.
Microsoft recently found that business email compromise (BEC), for instance, has increased significantly by 38% over the past four years, and a recent report from Osterman Research revealed that large organizations are expecting another 43% surge in the next 12 months.
Themis Co-pilot is currently available through a beta program.