All tracked items across vulnerabilities, news, research, incidents, and regulatory updates.
Network traffic patterns constantly change, causing traditional malicious traffic detection systems to become less effective over time, a problem called concept drift (when the patterns an AI learned on no longer match real-world data). Researchers developed Argus, a framework that automatically detects when traffic patterns shift, identifies new malicious patterns without human help, and continuously updates itself to maintain high detection accuracy even as attacks evolve.
According to Stanford economist Mordecai Kurz, tech billionaires are concentrating technological power in a way that weakens democracy, similar to patterns seen during industrialization and the first Gilded Age. Kurz argues that tech moguls often see themselves as superior beings meant to reshape society, citing examples like Anthropic's CEO claiming AI could become a transcendent good while potentially causing mass unemployment.
OpenAI and Dell Technologies are partnering to let businesses use Codex (an AI tool that writes and understands code) in their own private data centers and hybrid environments (networks that combine on-site systems with cloud services) rather than only in the cloud. This allows companies to keep sensitive data secure while using Codex across their existing tools and workflows for both coding tasks and broader business automation.
Organizations buy many security tools to address threats, but this approach fails because companies lack visibility, a unified understanding of their entire IT environment and what each tool is actually monitoring. Attackers exploit the gaps between tools by moving through systems using legitimate access rather than breaking through defenses, meaning the real security problem isn't inadequate tools but an incomplete map of what's happening across all systems.
This article covers a legal dispute between Elon Musk and Sam Altman over OpenAI's conversion from a nonprofit to a for-profit company. Musk founded OpenAI in 2015 with Altman to prevent Google from monopolizing AI technology, but sued in 2024 claiming they violated their commitment to keep it nonprofit, while Altman argues no such commitment was ever made. The article focuses on their trial testimony rather than any technical AI issue or security concern.
Google has introduced Gemini Omni Flash, a new AI model that can generate and edit videos from text, images, audio, or video inputs combined together. The model uses reasoning about physics and real-world knowledge to create realistic videos and allows users to edit them through natural language conversation (giving text instructions rather than using traditional editing tools), with changes building on each other while maintaining consistency in characters, physics, and scene details.
Face-swapping deepfakes (AI-generated videos or images where one person's face is replaced with another) are widely misused for fraud and misinformation, and while detection tools exist, there has been little work on tracing and recovering the original face that was replaced. This paper presents FaceReclaim, a new AI method that uses diffusion models (neural networks trained to gradually refine noisy images into clear ones) to restore the original face from a deepfaked image by separating facial attributes like expressions from identity information.
Backdoor attacks (hidden triggers that manipulate AI model predictions while keeping normal performance intact) are a serious security threat to deep neural networks (machine learning models with many layers). This paper presents PVDI, a defense method that removes backdoors by selectively preserving important attention patterns (the AI's focus on relevant input features) while disrupting irrelevant ones, successfully reducing attack success rates without hurting the model's normal performance.
Website fingerprinting (WF) attacks identify which websites users visit on Tor, a privacy network, but struggle when traffic patterns differ between training and real-world scenarios. This research presents UDA-WF, a new method using unsupervised domain adaptation (a machine learning technique that helps models work across different data distributions) to identify websites more efficiently with less training data. UDA-WF reduces the auxiliary data needed by 95% while maintaining 97.37% accuracy.
AI agents (autonomous programs that can perform tasks with minimal human direction) are becoming sophisticated enough to find and exploit obscure vulnerabilities (weaknesses in software), while at the same time developers are creating enormous amounts of AI-generated code that may contain bugs. This combination is forcing security teams to develop new defense strategies.
Researchers have developed CrossMPI, an image-based prompt injection attack (tricking an AI by hiding instructions in its input) that uses nearly invisible changes to images to manipulate how multimodal AI systems (AI that processes both images and text) interpret user instructions without modifying the text itself. The attack successfully fooled multiple vision-language models (AI systems that understand both images and text) about 66% of the time by targeting the intermediate layers where visual and textual information are combined, posing growing security risks as enterprises increasingly adopt multimodal AI systems.
AI-assisted coding is causing a rapid increase in leaked secrets (authentication credentials and API keys), with AI-related secrets exposed jumping 81% in 2025 alone, because developers prioritize speed and functionality over security reviews. When secrets are discovered, organizations should treat them as security incidents, immediately revoking or disabling the exposed credential, generating a new one, investigating system impact, performing cleanup, and hardening systems, followed by post-mortems to improve processes.
Fix: When a leaked secret is detected, organizations should: (1) activate their incident response process immediately; (2) revoke or disable the secret and generate a new one; (3) have the incident response team and R&D investigate the impact across systems and data; (4) perform cleanup and hardening; and (5) conduct post-mortems and implement necessary updates to systems or policies based on lessons learned. The source notes that the CISO office typically coordinates incidents while the R&D team owns actual revocation and cleanup.
CSO OnlineSouth Korea is using its upcoming local elections as a test case to see whether laws can effectively stop deepfakes (fake videos or audio created using AI to manipulate what people look like or sound like). The country is examining whether regulation can reduce the spread of these manipulated media during elections.
This research paper, published in May 2026, discusses a system that automatically chooses appropriate security protections to reduce risks in software applications. The work addresses how to match the right defensive techniques to specific vulnerabilities without requiring manual human selection.
This academic article examines how to secure business communications by combining human-focused practices with new research findings. The work suggests that protecting enterprise systems requires attention to both the people using them and technological solutions.
This study tested whether Western theories about why employees follow security policies apply to Saudi workers by surveying 401 employees. The research found that cultural differences and local policies significantly affect how employees think about security compliance, meaning that strategies to encourage safe behavior need to be tailored to specific cultures rather than using one-size-fits-all approaches.
This research examines why patients hesitate to share health information with health information technology systems (HIT, software that stores and manages medical records). The study found that patients are more willing to share information when they feel in control of how the technology is used, when they trust that their data is protected by security measures and regulations, and when they perceive real benefits from sharing. Conversely, patients become less willing to share when they feel their data is being tracked without their knowledge.
A study of 350 MBA students and 42 information systems graduates found that feedback from generative AI (AI systems that create new text or content) has mixed effects on workers: it boosts confidence in their abilities and motivation, but simultaneously makes them feel devalued and replaceable because the AI can perform the same tasks independently. The research also discovered that GenAI creates 'prompt engineering convergence' (where different types of work become repetitive prompting and reviewing tasks), which doesn't motivate workers the way traditional job variety does.
This report presents a framework for measuring research impact in Information Systems, a field where traditional academic metrics (like citation counts) don't capture the full value of research for organizations and society. Researchers from a global workshop developed a matrix that evaluates research across six themes (such as stakeholder engagement and ethics) and four project phases (planning, delivering, measuring, communicating) to help IS researchers design more impactful work.
Remote patient monitoring (RPM), a system using information and communication technologies to track patients' health from a distance, has expanded rapidly due to COVID-19 and payment policy changes, but faces significant challenges in how healthcare data is managed across fragmented systems. The main obstacles fall into three areas: trust and responsibility issues, limited and disconnected infrastructure (the technical systems that don't work well together), and changes in how healthcare workers do their jobs and what skills they need. The article calls for future research and curriculum changes to help information systems professionals address these challenges.