How FraudGPT presages the future of weaponized AI
VentureBeat: FraudGPT is a new generative AI tool found on the dark web, created to aid even novices in crafting malicious cyberattacks. It offers functionalities like creating malware, writing phishing emails, and discovering vulnerabilities for a subscription fee. While not as advanced as nation-state cyber units, its accessibility could lead to increased attacks on vulnerable sectors such as education and healthcare. The rise of weaponized generative AI marks a new phase in the cybersecurity arms race, necessitating robust AI-driven defenses and innovative red-teaming exercises to anticipate potential threats.
Data centers at risk due to flaws in power management software
CyberScoop: Data centers are increasingly vulnerable to hacker disruptions due to power management system flaws. At DEF CON, Trellix researchers highlighted vulnerabilities in popular data center applications, potentially allowing hackers access and the ability to cut power to specific servers. With the rise in cloud computing reliance, such software vulnerabilities can be catastrophic. The research pinpointed flaws in CyberPower’s management platform and Dataprobe’s units, which could cause significant business losses. Both companies have since patched these vulnerabilities.
Discord.io Temporarily Shuts Down Amid Breach Investigation
Dark Reading: Discord.io, a third-party service popular for sending Discord invites, suffered a data breach compromising information of 760,000 members. While the attacker remains unidentified, a vulnerability in Discord.io’s website code is believed to be the cause, enabling the threat actor to access and subsequently sell the database. Exposed data includes both sensitive and nonsensitive details, like usernames, email addresses, passwords, and more. No payment data was compromised. The platform has since halted operations, recommending password changes for users predating 2018. Discord.io is overhauling its security and plans a website code rewrite to prevent future breaches.
98% of organizations worldwide have integrations with at least one third-party vendor that has been breached in the last two years. (source)
With a 95% accuracy rate, a newly developed AI model discerned typing and stole passwords, using keystrokes transmitted from a laptop through a smartphone. (source)