AI Sec Watch

This research proposes K-TCDP (K-Temporal Correlated Differential Privacy), a new method for training large language models privately using LoRA (a technique that adds small trainable adapters to a model). Standard privacy-preserving training adds random noise that degrades model quality, but K-TCDP uses strategically correlated noise over time so that noise added in early steps can be partially canceled out by noise in later steps, improving model performance while maintaining privacy guarantees.

This research addresses privacy and data quality challenges in federated learning (FL, a technique where multiple computers train an AI model together without sharing raw data) for skeleton-based action recognition (identifying human movements from body joint positions). The authors propose Fed-C&E, a system that uses data condensation on client devices to reduce privacy risks, then expands the condensed data on a central server using techniques like a prototype-to-sequence similarity transformation matrix pool and feature expansion with second-order statistics to recover lost information and prevent overfitting.

When employees connect unapproved AI apps to work platforms like Google Workspace or Salesforce using OAuth (a system that lets apps access your accounts), they create persistent bridges that attackers can exploit if the AI app gets hacked. The Vercel breach showed this risk in action: an employee used a trial version of Context.ai without approval, and when Context.ai was compromised, attackers used the OAuth tokens (digital keys that grant access) to reach sensitive Vercel data like API keys and employee records.

Scammers are creating deepfakes (AI-generated fake videos that realistically mimic real people) of celebrities like Taylor Swift and Rihanna on TikTok to trick users into fake reward programs. These deepfakes often manipulate real footage with AI and use TikTok's official branding to appear legitimate, but they redirect users to third-party websites that steal personal information.

Ollama for Windows has a remote code execution vulnerability (the ability for an attacker to run commands on your computer) in its update system. The vulnerability happens because the application builds file paths using information from HTTP headers without checking if they're legitimate, allowing attackers to use path traversal sequences (like ../ to navigate directories) to write malicious executable files to dangerous locations like the Windows Startup folder. When combined with a missing signature verification flaw, an attacker can automatically execute malicious code without the user knowing.

Ollama for Windows has a vulnerability (CVE-2026-42248) where it does not verify that downloaded updates are authentic and haven't been tampered with before installing them. Because Ollama automatically installs updates without asking the user, an attacker could trick the software into downloading and running malicious code without the user knowing.

Threat actors are now using custom AI systems to automate cyberattacks, such as mapping Active Directory (a system that manages user accounts and permissions in networks) and stealing admin credentials within minutes, moving much faster than traditional security teams can respond. Traditional defense workflows involve multiple teams working in silos (separate, disconnected groups) with slow handoffs between threat intelligence, red team testing (simulated attacks to find weaknesses), and blue team patching (fixing vulnerabilities), creating dangerous delays. The webinar promotes "Autonomous Exposure Validation" as a new defensive approach to speed up security responses and eliminate these organizational bottlenecks.

OpenAI, a private company valued at over $850 billion, has become a major influence on tech earnings this week as four hyperscalers (Amazon, Alphabet, Meta, and Microsoft, the largest computing companies) report quarterly results. After a Wall Street Journal report suggested OpenAI missed revenue and user growth targets and may struggle to afford its data center expansion, investors are closely watching how this affects the companies that have invested billions in OpenAI or depend on its technology.

Firefox discovered 271 zero-day vulnerabilities (previously unknown security flaws) using Claude Mythos Preview, an advanced AI model from Anthropic, with fixes included in Firefox 150. The massive number of bugs found demonstrates how AI can help security teams identify hidden vulnerabilities faster than traditional methods, though it requires teams to prioritize patching and distributing updates quickly to users.