Applied GAI in Security

Sign in Subscribe

Brandon Dixon

Brandon Dixon

Brandon is a cybersecurity expert currently leading Security Copilot at Microsoft. He is known for creating PassiveTotal, Blockade.io, NinjaJobs, PDF X-Ray, and other opensource tools

Last Week in GAI Security Research - 07/01/24

Last Week in GAI Security Research - 07/01/24

Highlights from Last Week * 🪱 Synthetic Cancer – Augmenting Worms with LLMs * 🔗 Large Language Models for Link Stealing Attacks Against Graph Neural Networks * 🧑‍💻 Assessing the Effectiveness of LLMs in Android Application Vulnerability Analysis * 🦠 MALSIGHT: Exploring Malicious Source Code and Benign Pseudocode for Iterative Binary Malware Summarization * 🦜 Poisoned LangChain: Jailbreak LLMs by LangChain

Last Week in GAI Security Research - 06/24/24

Last Week in GAI Security Research - 06/24/24

Highlights from Last Week * 🥷🏿 AgentDojo: A Dynamic Environment to Evaluate Attacks and Defenses for LLM Agents * 🐵 From LLMs to MLLMs: Exploring the Landscape of Multimodal Jailbreaking * 💰 Jailbreaking as a Reward Misspecification Problem * 🏹 "Not Aligned" is Not "Malicious": Being Careful about Hallucinations of Large Language Models'

Last Week in GAI Security Research - 06/17/24

Last Week in GAI Security Research - 06/17/24

Highlights from Last Week * 🛍 Bag of Tricks: Benchmarking of Jailbreak Attacks on LLMs * 🕵‍♀ Security Vulnerability Detection with Multitask Self-Instructed Fine-Tuning of Large Language Models * 🚪 A Survey of Backdoor Attacks and Defenses on Large Language Models: Implications for Security Measures * 🤖 Machine Against the RAG: Jamming Retrieval-Augmented Generation with

Last Week in GAI Security Research - 06/10/24

Last Week in GAI Security Research - 06/10/24

Highlights from Last Week * 💉Exfiltration of personal information from ChatGPT via prompt injection * 🦥 Are you still on track!? Catching LLM Task Drift with Activations * 🕵 Ranking Manipulation for Conversational Search Engines * 💔 BadRAG: Identifying Vulnerabilities in Retrieval Augmented Generation of Large Language Models * 🎼 Transforming Computer Security and Public Trust Through the Exploration

Automating Incident Triage with Copilot for Security

Automating Incident Triage with Copilot for Security

When speaking with Copilot for Security customers, automation is often brought up as a topic of exploration. Customers are eager to extend their existing SOAR investments or workflows to include Copilot because they recognize the capabilities this new technology brings and believe it has the potential to further increase productivity.

Last Week in GAI Security Research - 06/03/24

Last Week in GAI Security Research - 06/03/24

Highlights from Last Week * 👹 The Peripatetic Hater: Predicting Movement Among Hate Subreddits * ⌛ Exploiting LLM Quantization * 📈 Intruding with Words: Towards Understanding Graph Injection Attacks at the Text Level * 🛡️ Genshin: General Shield for Natural Language Processing with Large Language Models * 🌤️ Unlearning Climate Misinformation in Large Language Models Partner Content Codemod is the

Last Week in GAI Security Research - 05/27/24

Last Week in GAI Security Research - 05/27/24

Highlights from Last Week * 💡 Generative AI and Large Language Models for Cyber Security: All Insights You Need * 🌐 Impact of Non-Standard Unicode Characters on Security and Comprehension in Large Language Models * 🛜 Large Language Models in Wireless Application Design: In-Context Learning-enhanced Automatic Network Intrusion Detection * 👓 Towards Transferable Attacks Against