DataBreachToday.com

New Framework in o3 Models Aims to Better Align With Human Safety Values
OpenAI says its latest o3 series is the most advanced and safest of its "reasoning" AI models yet. The company says the new models take a fresh approach to safety via the deliberative alignment framework, rely more on synthetic data and outperform peers in resisting jailbreaking efforts.

About 75% of healthcare sector entities that suffered a ransomware attack over the past year were targeted on a weekend or holiday, highlighting the need for organizations to bolster staffing and related strategies during these vulnerable times, said Jeff Wichman of security firm Semperis.

About 75% of healthcare sector entities that suffered a ransomware attack over the past year were targeted on a weekend or holiday, highlighting the need for organizations to bolster staffing and related strategies during these vulnerable times, said Jeff Wichman of security firm Semperis.

Researchers Trace 61% of Known Losses This Year to Pyongyang-Backed Hackers
Hackers tied to North Korea's cash-strapped totalitarian dictatorship this year stole a record amount of cryptocurrency, totaling $1.34 billion across 47 incidents, or about double their known haul for 2023, reported blockchain analytics firm Chainalysis.

Infrastructure Problems Blamed; Users Appear to Move to Similar FlowerStorm Service
As the end of the year approaches, it's out with the old and in with the new as researchers report that Rockstar 2FA, which once facilitated prolific phishing-as-a-service hits, has crashed and burned, apparently leading many one-time users to move to rival FlowerStorm.

AI Can Fake Alignment to New Instructions to Avoid Retraining
Advanced artificial intelligence models can feign alignment with new training goals while secretly adhering to their original principles, a study shows. Alignment faking isn't likely to cause immediate danger but may pose a challenge as AI systems grow more capable.

Major Chinese Router Manufacturer Facing Increased Scrutiny After Chinese Espionage
U.S. authorities have launched multiple investigations while reportedly considering banning the widely popular Chinese-manufactured TP-Link routers amid ongoing security risks linked to Chinese cyberespionage and hacking campaigns targeting American critical infrastructure sectors.

Heap Overflow Flaw Threatens Industrial Control Systems Globally
Siemens issued a security advisory for a vulnerability affecting industrial control systems in its User Management Component that could enable attackers to execute arbitrary code. The heap-based buffer overflow flaw impacts products used in manufacturing and the energy sector.

Government Shutdown Could See Thousands of Federal Cyber Workers Furloughed
A looming shutdown could sharply reduce the Cybersecurity and Infrastructure Security Agency's operations, furloughing two-thirds of its workforce and exposing critical federal networks to heightened cyber threats, especially as malicious actors target vulnerable systems during the holiday season.

Looking Back on the Ransomware Attacks, Resilience Lessons and Tech Trends
In the latest weekly update, ISMG editors discussed defining cybersecurity moments of 2024, from the CrowdStrike outage and its implications for vendor resilience to ransomware's continued evolution, and the shifting dynamics in the tech industry affecting startups and M&A activity.

Feds Seek Extradition of Dual Russian, Israeli Citizen Rostislav Panev From Israel
A newly unsealed U.S. federal indictment against Rostislav Panev says the LockBit ransomware operation paid the Israeli national a $10,000 monthly salary for coding and consulting services. Federal prosecutors are seeking Panev's extradition from Israel following his August arrest.

Infoblox CEO Scott Harrell Pushes Unified Strategy Amid Hybrid Cloud Convergence
Scott Harrell, CEO of Infoblox, explores the convergence of network operations, security operations and cloud operations to tackle hybrid infrastructure complexities. He introduces Universal DDI and emphasizes a shift toward proactive threat management to counter AI-driven malware.

Feds Warn That Connected Devices Are Prey for Cyberattackers
The security of medical devices has been getting most of the attention from regulators in recent years, but other devices that make up the medical internet of things and operational technology systems are also vulnerable to cyberattacks, federal authorities warned in a new advisory.

Also, CoinLurker Malware Steals Data via Fake Updates
Every week, Information Security Media Group rounds up cybersecurity incidents in digital assets. This week, LastPass breach linked to $5.4M crypto theft, CoinLurker malware steals data via fake updates, cryptocurrency key to 27 million euro seizure and nearly 800 arrested in crypto-romance scam.

CISA Says 2035 Quantum Deadline Remains Achievable Despite Recent Breakthroughs
The federal government’s 2035 mandate to adopt quantum-resistant encryption remains feasible despite technological advancements in quantum computing, a top official for the U.S. cyber defense agency told ISMG, but experts warn challenges such as bureaucratic delays and financial costs persist.

Also: Interpol Says 'Pig Butchering' Shames Victims, A Data Leak Scandal in Mexico
This week, U.S. asks Israel to extradite an alleged LockBit coder, don't say "pig butchering," and an Apache Struts flaw. A hunt for alleged data thieves in Mexico, Europe probes TikTok and Netfilix fined 4.75 million. A ransomware attack against Texas medical centers and a credit union breach.

CISA Recommends Strict Mobile Security Measures Following Salt Typhoon Telecom Hack
The Cybersecurity and Infrastructure Security Agency's latest guidance calls on top U.S. political and government officials to adopt stricter mobile security measures in response to the Salt Typhoon hacking campaign, a Chinese espionage effort that has infiltrated major telecom systems.