News
Latest cybersecurity news — hacks, breaches, vulnerabilities, regulatory moves
Encryptionless Extortion: Why Ransomware Stopped Encrypting and Started Leaking
More crews now steal and threaten to leak instead of encrypting. Your backups will not save you from this one.
NewsEDR Killers and BYOVD: How 2026 Ransomware Disables Your Defences First
Before the ransomware fires, the operator kills your EDR. Here is how BYOVD works and how to block it.
AI SecuritySecuring the AI Supply Chain in 2026: Model Provenance, Poisoning, and MCP Tool Risk
Your AI is only as trustworthy as the weights, data, and tools behind it. The 2026 guide to AI supply-chain risk.
NewsOpen-Source Reasoning Models Closed the Gap — and On-Prem AI Just Got Real
Open-source models are now serious. For regulated Indian sectors, local deployment is finally a real option — with its own security trade-offs.
AI SecurityAI Governance in 2026: Early Model Access for Regulators, the EU AI Act, and DPDP × AI
Governments are pushing pre-release testing and early model access. For Indian teams, AI governance now sits at the intersection of DPDP and…
NewsEnterprise AI in 2026: From Pilot Project to Core Infrastructure
JPMorgan reclassified AI as core infrastructure; labs are embedding engineers inside businesses. AI just became a system you have to secure like…
AI SecurityShadow AI Is Your Biggest Data-Leak Risk in 2026 — and DPDP Makes It Expensive
Unsanctioned AI tools are now the #1 data-leak entry point. Under the DPDP Act, that is not just a security problem —…
AI SecurityAI Red-Teaming Goes Autonomous: Agents That Hack Other AIs
Autonomous agents now solve most black-box red-team challenges faster than humans. What that means for offensive security in 2026.
NewsThe 2026 Frontier Model Landscape: GPT-5.4, Claude 4.6, Gemini 3.1 and Llama 4
GPT-5.4, Claude 4.6, Gemini 3.1 and Llama 4 reset the baseline in 2026. What changed, and what it means if you build…
AI SecuritySecond-Order Prompt Injection: How Attackers Hijack Multi-Agent Systems
A new attack class: feed a low-privilege agent a malformed request so it convinces a higher-privilege agent to act. Here is how…