Cybersecurity researchers are sounding the alarm on a dual-pronged threat targeting Brazil. In one campaign, threat actors are leveraging legitimate generative AI tools to create highly convincing phishing pages of Brazilian government agencies to trick users into making payments. These fraudulent sites are boosted with SEO poisoning to appear in top search results. Simultaneously, a separate malspam campaign is distributing the Efimer trojan, a potent malware designed to steal cryptocurrency, which has already impacted over 5,000 users.
A massive, AI-driven scam campaign codenamed 'ClickTok' is targeting TikTok Shop users worldwide. Cybersecurity firm CTM360 reports that threat actors have created over 15,000 fake TikTok Shop domains to execute a dual-pronged attack involving phishing and malware distribution. The campaign leverages AI-generated videos and fake influencer promotions on Meta platforms to lure victims to these lookalike sites. Once there, users are tricked into entering credentials on phishing pages or downloading a trojanized TikTok app. This malicious app contains the SparkKitty malware, a cross-platform threat designed to steal sensitive data, including cryptocurrency wallet seed phrases from screenshots on both Android and iOS devices. The scam's primary goals are financial, using fake product listings, fraudulent crypto payments, and credential theft to exploit both buyers and affiliate sellers on the popular e-commerce platform.
A critical vulnerability, CVE-2025-23266, in NVIDIA's Container Toolkit allows for complete server takeover on shared AI cloud services with a simple exploit.
Security researchers have discovered a serious vulnerability in Google Gemini for Workspace that allows attackers to inject hidden commands into emails, leading to convincing phishing attempts—all without links or attachments.
A major security flaw in McDonald's AI hiring system, McHire, exposed personal data of 64 million job applicants due to a weak password. McDonald's AI Hiring Bot Leaks 64M Applicant Data
