Cyber Due Diligence
It Takes Only 250 Documents to Poison Any AI Model
Researchers find it takes far less to manipulate a large language model's (LLM) behavior than anyone previously assumed.
Too Many Secrets: Attackers Pounce on Sensitive Data Sprawl
Hardcoded credentials, access tokens, and API keys are ending up in the darnedest places, prompting a call for organizations to stop ov...
WhatsApp Secures Ban on NSO Group After 6-Year Legal Battle
NSO Group must pay $4 million in damages and is permanently prohibited from reverse-engineering WhatsApp or creating new accounts after...
Russia Pivots, Cracks Down on Resident Hackers
Thanks to improving cybersecurity and law enforcement action from the West, Russia's government is reevaluating which cybercriminals it...
MuddyWater Targets 100+ Gov Entities in MEA With Phoenix Backdoor
The Iranian threat group is using a compromised mailbox accessed through NordVPN to send phishing emails that prompt recipients to enab...
Verizon: Mobile Blindspot Leads to Needless Data Breaches
People habitually ignore cybersecurity on their phones. Instead of compensating for that, organizations are falling into the very same ...
Asian Nations Ramp Up Pressure on Cybercrime ‘Scam Factories’
After a particularly gruesome murder, South Korea issues "code black" travel ban for several regions in Cambodia, while other nations u...
Electronic Warfare Puts Commercial GPS Users on Notice
Interference with the global positioning system (GPS) isn't just a problem for airlines, but for shipping, trucking, car navigation, ag...
Streaming Fraud Campaigns Rely on AI Tools, Bots
Fraudsters are using generative AI to generate fake music and boost the popularity of the fake content.