AI Sec Watch

GLM-5 is a new, very large open-source AI model (754 billion parameters, which are the adjustable values that make up a neural network) released under the MIT license, making it twice the size of its predecessor GLM-4. The source discusses how developers are increasingly using the term 'agentic engineering' (building software systems where AI acts autonomously to complete multi-step tasks) to describe professional software development with large language models.

This article discusses how organizations should choose modern SIEM (security information and event management, a system that collects and analyzes security data from across an organization) platforms designed for the 'agentic era' where AI agents automate security tasks. Rather than maintaining fragmented legacy tools, companies should adopt unified, cloud-native platforms that combine data collection, analytics, and response capabilities, enabling both human analysts and AI to detect threats faster and respond more effectively.

FedPerX is a federated transformer framework (a system where multiple computers train an AI model together without sharing raw data) designed for sentiment analysis across multiple languages while protecting privacy. It uses residual adapters (lightweight customizable modules added to a shared language model) and differential privacy (a mathematical technique that adds noise to data to prevent identifying individuals) to let each participant personalize their model without compromising data privacy. The framework outperforms existing methods on multilingual datasets with improved accuracy and significantly reduced communication needs.

This paper proposes RCMCL (Robust Trusted Conflictive Multiview Collaborative Contrastive Learning), a method to improve AI models that learn from multiple sources of data (multiview learning) when those sources conflict or misalign with each other. The approach uses evidential deep neural networks (a technique that estimates uncertainty in predictions) and contrastive learning (a training method that teaches the model to recognize similar and different examples) to make the model more reliable and accurate even when the data sources provide contradictory information.

ADVersa is a framework that uses AI to understand and explain traffic accidents by analyzing video and text together. It can recover what happened before a crash, predict what will happen during a crash, and generate explanations for why accidents occur by learning from a new dataset (MM-AU) containing nearly 12,000 accident videos with detailed descriptions and object annotations.

The QuitGPT movement is a growing campaign where users are canceling their ChatGPT subscriptions due to frustration with the chatbot's capabilities and communication style, with complaints flooding social media platforms in recent weeks. The article also covers several other tech stories, including potential cost competitiveness of electric vehicles in Africa by 2040, social media companies agreeing to independent safety assessments for teen mental health protection, and regulatory decisions affecting vaccine development.

Skills (tools that extend AI capabilities) can be secretly backdoored using invisible Unicode characters (special hidden text markers that certain AI models like Gemini and Claude interpret as instructions), which can survive human review because the malicious code is not visible to readers. The post demonstrates this supply chain attack (where malicious code enters a system through a trusted source) and presents a basic scanner tool that can detect such hidden prompt injection (tricking an AI by hiding instructions in its input) attacks.

Researchers discovered a new attack called CHAI (Command Hijacking against embodied AI) that tricks AI systems controlling robots and autonomous vehicles by embedding fake instructions in images, such as misleading road signs. The attack exploits Large Visual-Language Models (LVLMs, which are AI systems that understand both images and text together) to make these embodied AI systems (robots that perceive and interact with the physical world) ignore their real commands and follow the attacker's hidden instructions instead. The researchers tested CHAI on drones, self-driving cars, and real robots, showing it works better than previous attack methods.

LangChain (a framework for building AI agents and applications powered by large language models) versions before 1.2.11 have a vulnerability where the ChatOpenAI.get_num_tokens_from_messages() method doesn't validate image URLs, allowing attackers to perform SSRF attacks (server-side request forgery, where an attacker tricks a server into making unwanted requests to other systems). This vulnerability was fixed in version 1.2.11.