“We’ll have a generation of security professionals who can supervise AI but can’t function without it."

Categories: AI Research, Sophos Insights

Tags: AI, AI Cybersecurity, AI RESEARCH, Generative AI, SOC

Following on from our preview, here’s the full rundown on LLM salting: a novel countermeasure against LLM jailbreaks, developed by AI researchers at Sophos X-Ops

Categories: AI Research

Tags: AI, CAMLIS, Featured, jailbreak, LLM, salting, Sophos X-Ops

On October 22-24, SophosAI will present research on ‘LLM salting’ (a novel countermeasure against jailbreaks) and command line classification at CAMLIS 2025

Categories: AI Research

Tags: AI, CAMLIS, Featured, LLM, Sophos X-Ops

Analyzing dark web forums to identify key experts on e-crime

Categories: AI Research, Threat Research

Tags: AI, cybercrime, Dark Web, Featured, threat activity cluster, threat actors

Sophos X-Ops’ research, presented at Virus Bulletin 2024, uses ‘multimodal’ AI to classify spam, phishing, and unsafe web content

Categories: AI Research

Tags: Featured, Large Language Models, Multimodal AI, Sophos X-Ops, spam detection, Web Content Filtering

SophosAI’s framework for upgrading the performance of LLMs for cybersecurity tasks (or any other specific task) is now open source.

Categories: AI Research

Tags: deepspeed, Featured, LLM, LLM tuning

“LLMbotomy” research reveals how Trojans can be injected into Large Language Models, and how to disarm them.

Categories: AI Research

Tags: AI Trojans, Featured, LLM

On October 24 and 25, SophosAI presents ideas on how to use models large and small—and defend against malignant ones.

Categories: AI Research

Tags: AI Trojans, anti-phishing, CAMLIS, Featured, Google, LLM, small model machine learning

Applying generative AI, bad actors could tailor disinformation campaigns to affect election outcomes on a massive scale with relatively little effort.

Categories: AI Research

Tags: adversarial ai, Featured, Generative AI, misinformation, scampaign

Sophos' Younghoo Lee will present his research on the use of AI to analyze both text and image data to classify spam, phishing, and unsafe web content in Dublin.

Categories: AI Research

Tags: anti-phishing, Featured, Large Language Models, Multimodal AI, spam detection, Web Content Filtering

Comparative Sophos X-Ops testing not only indicates which models fare best in cybersecurity, but where cybersecurity fares best in AI

Categories: AI Research

Tags: Featured, Large Language Models

Categories: AI Research, Threat Research

Tags: adversarial ai, artificial intelligence, Featured, Generative AI, scams, Sophos X-Ops

The conference on machine learning in cybersecurity is key to open exchange of research and knowledge.

Categories: AI Research

Tags: artificial intelligence, CAMLIS, Featured, Large Language Models, scams, Web Content Filtering

AI Village talk highlights how generative  can be used to automate the creation of fraud campaigns, generating hundreds of fraudulent sites.

Categories: AI Research

Tags: adversarial ai, DEF CON, Generative AI, Large Language Models, web scams

Sophos AI team employs GPT and other large language models as teachers to train smaller models to label websites.

Categories: AI Research

Tags: BERT, Featured, GPT-3, Large Language Models, Sophos X-Ops, T5 Large LLM, Web filtering, website categorization