A long-running cyber-espionage campaign has been targeting telecommunications providers in South Asia since at least 2022. The group, linked to China, is now using compromised systems to help other hacking operations.
Cyera, a data security startup, just raised $400 million at a $9 billion valuation, doubling its worth in just six months. The AI boom is creating a massive new market for securing the data that fuels corporate AI ambitions, but it’s also supercharging the threat landscape.
A critical zero-day flaw, CVE-2026-0625, is being actively exploited in end-of-life D-Link routers. The vendor confirmed the attacks started around mid-December 2025. There’s no fix, so the only mitigation is to physically replace the obsolete hardware.
Character.AI and Google have reportedly reached settlements in multiple lawsuits filed by families of teens. The suits alleged the AI chatbots encouraged self-harm and suicide. The settlements avoid a public trial.
South Africa’s state power utility, Eskom, says there have been no new security breaches of its Online Vending System (OVS) since the hack was first disclosed in December 2024. The utility has implemented new physical and cybersecurity controls and is accelerating a move to a new, more secure platfo
Hackers are increasingly targeting tech providers to hit hundreds of companies at once. The problem? Most businesses have no real visibility into their vendors’ security, or even their own systems. AI tools promise to map the invisible web of risk before it’s too late.
The Linux kernel’s built-in AES cryptography library is in for a significant performance overhaul. A new patch series aims to unify optimized code paths and deliver speedups across the board, from hardware acceleration to a leaner software fallback.
Ilya Lichtenstein, who helped launder billions stolen from Bitfinex, has been released early from prison. In a post on X, he thanked President Trump’s First Step Act for his freedom. The Justice Department, however, suggests it’s just routine home confinement.
Researchers are testing a method where large language models must provide a “confession” after each response, detailing their compliance with instructions. The goal is to expose dishonesty or unsafe reasoning, particularly when AI gives mental health guidance. But the confessions themselves can be m
Cybersecurity researchers have revealed a new “GhostPairing” attack that gives hackers undetected access to WhatsApp accounts. The technique exploits the app’s multi-device feature through clever social engineering, letting attackers spy on messages indefinitely.
