Security | Threat Detection | Cyberattacks | DevSecOps | Compliance

Language Switching Attacks: The New Threat Vector in LLM Security In this clip from "Securing AI Part 4: The Rising Threat of Hidden Attacks in Multimodal AI," Diptanshu Purwar discusses the growing trend of language-switching attacks. These techniques exploit the ongoing development and training gaps in Large Language Models (LLMs). Diptanshu explains how attackers can evade an LLM's built-in filters and guardrails by rapidly shifting between different languages, particularly less common ones, to find weaknesses where the model's safety data is sparse.

Something unexplained has appeared inside Securonix. No sender. No clue. Just a steady glow that feels… aware. It doesn’t open. It doesn’t move. But everyone can sense it’s waiting for us to notice. The mystery begins here.

New question. Real answers. No fluff. Welcome back to Off the Blocks — Fireblocks’ rapid-fire video series, shot live at TOKEN2049 Singapore. In Episode 3, we asked industry leaders just one thing: In one sentence, what does the future of onchain finance mean to you? From programmable liquidity to permissioned DeFi, their responses are bold, honest, and sharply focused on what comes next. This is where ideas become infrastructure, and where vision meets execution.

In this video, we dive into Cursor 2.0 and Composer. Watch as we write, refine, and analyze code in real time, then test how secure the generated code really is...

CrowdStrike Falcon Exposure Management simplifies network security with its Network Vulnerability Assessment, offering a modern way to secure network infrastructure without complex legacy tools. Learn how the system utilizes your existing Falcon sensors to provide unmatched continuous visibility-instantly correlating new threats against your assets—and leverage authenticated scanning for deeper, more complete vulnerability inspection.

In this episode of The Cybersecurity Defenders Podcast, we discuss some intel being shared in the LimaCharlie community. Support our show by sharing your favorite episodes with a friend, subscribe, give us a rating or leave a comment on your podcast platform. This podcast is brought to you by LimaCharlie, maker of the SecOps Cloud Platform, infrastructure for SecOps where everything is built API first. Scale with confidence as your business grows.

Microsoft’s DART team identified a new backdoor named SESAMEOP in July 2025 that uses the OpenAI Assistants API as its command and control (C2) channel. Proofpoint has detailed a campaign targeting freight and trucking companies using remote monitoring and management (RMM) tools to steal cargo. Security researchers at Catchify identified a critical unauthenticated remote code execution vulnerability in the UniFi Access backup/export workflow, tracked as CVE-2025-52665, with a CVSS score of 10.0 (critical).

How do you explain AI Security without the jargon? Easy you make it food. In this video, we asked leading AI Security professionals to describe AI Security as a dish. Their answers turn complex ideas like prompt injection, data leaks, and model hardening into bite-sized insights you’ll actually remember. From layered lasagna to spicy tacos, each response brings a fresh perspective on what it means to build and protect secure AI systems.

In this session, Or Amir, Product Manager at Kovrr, showcases our new AI Risk Assessment and AI Risk Quantification modules — helping enterprises gain visibility, benchmark maturity, identify shadow AI, and turn exposure into measurable outcomes.