Daven Tamayo

Fake ChatGPT, Claude, and Gemini apps fuel surge in global malware attacks—Kaspersky

Article (news)

THUMBNAIL-CYBER

Arthur Fuentes

ABS-CBN News

MANILA — Cybercriminals are increasingly exploiting the artificial intelligence boom, with security firm Kaspersky detecting over 92,000 malware attacks disguised as popular AI services globally in the first four months of 2026 alone.  Kaspersky said attackers are leveraging trusted brands to trick users into downloading malicious files. Fake versions of OpenAI’s ChatGPT accounted for nearly half (49 percent) of all detected attacks, while impersonations of Anthropic’s Claude and Google’s Gemini each made up 18 percent.  Since January, researchers identified more than 15,000 malware samples masquerading as agentic AI software, including counterfeit versions of rapidly growing open-source tools like OpenClaw. The malicious payloads included banking trojans, spyware, exploits, and downloaders.  Kaspersky said its Global Research and Analysis Team (GReAT) also uncovered a sophisticated campaign in May linked to the &quot;Silver Fox&quot; advanced persistent threat (APT) group. The operation distributed fake Claude AI applications for Windows, macOS, and Linux to silently install malware, granting attackers long-term access to sensitive information and compromised systems.  “The introduction of AI agents into enterprise environments changes the nature of trust itself. Every automated action becomes part of a wider chain of systems and data exchanges, which means security is no longer just about protecting endpoints – it is about controlling how intelligence, permissions, and decisions propagate across interconnected AI-driven processes,” said Dmitry Galov, Head of Russia and CIS units at Kaspersky GReAT.  To counter these evolving threats, Kaspersky urged organizations to deploy real-time threat visibility solutions and equip security teams with deep threat intelligence. Individual users were advised to stick to reputable AI providers, avoid anonymous bots designed to harvest data, and use robust security software to block phishing sites and malware installations.

Since January, researchers identified more than 15,000 malware samples masquerading as agentic AI software, including counterfeit versions of rapidly growing open-source tools like OpenClaw.

Since January, researchers identified more than 15,000 malware samples masquerading as agentic AI software, including counterfeit versions of rapidly growing open-source tools like OpenClaw.