AI tools and security flaws: a closer look at recent cyber threats

A group of curious internet users found a way into a restricted AI tool called Mythos Preview. They didn’t use advanced hacking skills—just careful detective work. After studying a breach at an AI training company, they guessed where Mythos was hosted and accessed it. They even found other unreleased AI models. So far, they’ve only used it for harmless projects, but their discovery shows how hard it is to truly lock down powerful AI tools. Meanwhile, cybersecurity experts warn that old phone network flaws are still being used for spying. A digital rights group found two companies abusing weaknesses in telecom systems to track people’s locations. These firms exploited small telecom providers to spy on targets, proving that global phone networks remain dangerously vulnerable.

Scammers are also getting bolder. US officials charged two men with running a fraud ring in Southeast Asia. They lured victims with fake job offers, trapped them in compounds, and forced them to scam people out of millions. The operation used crypto fraud and even involved physical punishment. Authorities froze $700 million in funds and seized a Telegram channel used to recruit victims. Another unsettling case involves health data. Three research groups in the UK were caught selling private medical records on Alibaba. Over 500, 000 people had trusted UK Biobank with their genetic and medical data, but some researchers broke the rules. The breach included sensitive details, raising serious questions about data protection. Even tech giants like Apple aren’t immune. A flaw in iPhones let the FBI access deleted Signal messages from a defendant’s device. While Signal’s encryption kept messages private, iOS saved copies in push notifications. Apple fixed this in an update, but the incident shows how small bugs can lead to big privacy risks.

Actions