Google has released VaultGemma, a large language model designed to keep sensitive data private during training. The model uses differential privacy techniques to prevent individual data points from being exposed, which makes it safer for handling confidential information in sectors…
Category: Help Net Security
Building security that protects customers, not just auditors
In this Help Net Security interview, Nir Rothenberg, CISO at Rapyd, discusses global differences in payment security maturity and the lessons that can be learned from leading regions. He points out that good engineering usually leads to strong security, and…
AI video surveillance could end privacy as we know it
AI-powered video surveillance brings up big questions about privacy. On one hand, it can make us feel safer, but on the other, it can easily cross the line into intrusion. The more we let technology watch and track our behavior,…
OT security needs continuous operations, not one-time fixes
Cyberattacks keep hitting the OT systems that critical infrastructure operators run, according to new research from Forrester. In a survey of 262 OT security decision-makers, 91% reported at least one breach or system failure caused by a cyberattack in the…
Product showcase: Clean Links exposes what’s hiding behind a QR code
Clean Links is a handy app that shows you exactly where a link will take you before you click it. It strips out trackers, expands shortened URLs, and helps you avoid scams while saving you time and frustration. The best…
Cybersecurity jobs available right now: September 16, 2025
CISO Haier Europe | Italy | On-site – View job details As a CISO, you will develop an information security strategy aligned with organizational priorities, secure executive support, and oversee the protection of data, intellectual property, and technology assets. You…
Phishing campaign targets Rust developers
Developers publishing crates (binaries and libraries written in Rust) on crates.io, Rust’s main public package registry, have been targeted with emails echoing the recent npm phishing campaign. The phishing email The emails started hitting developers’ inboxes on Friday, minutes after…
SecurityScorecard acquires HyperComply to automate vendor security reviews
SecurityScorecard announced the acquisition of HyperComply. Combining SecurityScorecard’s trusted security ratings and continuous third- and fourth-party risk visibility with HyperComply’s intelligent automation capabilities will make it possible for customers to stop the slow, manual process of conducting vendor security reviews…
Proofpoint launches agentic AI to detect risks in communication channels
Proofpoint launched agentic AI solution for Human Communications Intelligence (HCI), marking a leap forward in how organizations detect, understand, and mitigate conduct and compliance risks in real time. Designed for enterprises in regulated and highly litigious industries, it transforms digital…
Most enterprise AI use is invisible to security teams
Most enterprise AI activity is happening without the knowledge of IT and security teams. According to Lanai, 89% of AI use inside organizations goes unseen, creating risks around data privacy, compliance, and governance. This blind spot is growing as AI…