Perception Point, a leading provider of advanced threat prevention across digital communication channels, today reveals its latest detection innovation, developed to counter the emergent wave of AI-generated email threats. The AI-powered technology leverages Large Language Models (LLMs) and Deep Learning architecture to effectively detect and prevent Business Email Compromise (BEC) attacks, a cyber threat which is currently undergoing a seismic shift due to the rise of Generative AI (GenAI) technologies.
Threat actors are increasingly abusing evolving GenAI technology to perpetrate more sophisticated, highly targeted attacks against organizations of all sizes. Newly democratized capabilities, including the creation of high-quality, human-like outputs such as personalized emails, have transformed GenAI into a potent tool for cybercrime, particularly in the realm of social engineering and BEC attacks, which have scaled to new heights over the past year. BEC accounted for over 50% of incidents involving social engineering according to the DBIR 2023 Report, while Perception Point’s 2023 Annual Report revealed an 83% growth in BEC attempts.
In response to this escalating threat, Perception Point has developed an LLM-based detection model, innovatively harnessing the power of Transformers, AI models capable of understanding the semantic context of text – mirroring the technology behind popular LLMs like OpenAI’s ChatGPT and Google’s Bard. This approach enables Perception Point’s solution to identify the unique patterns in LLM-generated text, a critical factor in detecting and thwarting GenAI-based threats – a task traditional security vendors fail to achieve using contextual and behavioral analysis.
The model processes incoming emails at an average of 0.06 seconds, aligning with Perception Point’s ability to dynamically scan 100% of content in near real-time. It has initially been trained on hundreds of thousands of malicious samples caught by Perception Point, and is continuously updated with new data to maximize its effectiveness.
“Amid an increasingly complex threat landscape, there is an urgent need for cutting-edge defenses against GenAI-powered threats,” said Tal Zamir, CTO of Perception Point. “We’re being challenged as an industry with yet another avenue that bad actors have come to exploit in their ever-expanding range of attacks. By reversing this dynamic and proactively leveraging AI for detection, we are able to prevent these threats before they even reach the user’s inbox – a paradigm shift in the fight against BEC attacks.”
To minimize false positives, resulting from the widespread use of generative AI for crafting legitimate emails, Perception Point utilizes a unique 3-phase architecture. After initial scoring, the model categorizes the email content using Transformers and clustering algorithms, integrating insights from these steps with additional data such as sender reputation and authentication protocol information. This process allows the model to accurately predict whether the email is AI-generated and if it potentially poses a threat.
As GenAI continues to rise, Perception Point is leading the charge with its AI-powered cybersecurity solution. By leveraging the patterns in LLM-generated content, advanced image recognition, anti-evasion algorithms and patented dynamic engines, Perception Point provides a robust and proactive defense, neutralizing threats before they ever reach the user.