Skip to content
  • Home
  • Cyber Map
  • About Us – Contact
  • Disclaimer
  • Terms and Rules
  • Privacy Policy
Cyber Web Spider Blog – News

Cyber Web Spider Blog – News

Globe Threat Map provides a real-time, interactive 3D visualization of global cyber threats. Monitor DDoS attacks, malware, and hacking attempts with geo-located arcs on a rotating globe. Stay informed with live logs and archive stats.

  • Home
  • Cyber Map
  • Cyber Security News
  • Security Week News
  • The Hacker News
  • How To?
  • Toggle search form
New Android Malware Uses AI for Persistent Threats

New Android Malware Uses AI for Persistent Threats

Posted on February 19, 2026 By CWS

Cybersecurity experts have uncovered a new form of Android malware that exploits Google’s Gemini AI chatbot to establish and maintain its presence on devices. Known as PromptSpy, this malware employs advanced techniques to capture sensitive data and prevent its removal, making it a significant threat.

How PromptSpy Utilizes AI for Persistence

PromptSpy, identified by ESET researchers, leverages Gemini AI to analyze device screens and provide instructions to ensure its continued operation. This method prevents the malware from being easily closed or terminated by the system. By integrating AI into its processes, PromptSpy can adapt to various device configurations and operating systems, broadening its range of potential targets.

The malware hardcodes an AI model and prompts within its code, transforming Gemini into an ‘Android automation assistant.’ It sends a detailed XML dump of the current screen to Gemini, which responds with directions on actions to take, such as where and how to perform specific interactions. This process locks the malware into the recent apps list, making it persistent.

Technical Capabilities and Threats

PromptSpy’s primary function is to deploy a VNC module, allowing attackers remote access to infected devices. It exploits Android’s accessibility services to resist uninstallation and communicates with a command-and-control server using the VNC protocol. The malware can intercept lockscreen credentials, record screen activity, and communicate with a server to receive instructions and API keys.

Researchers observed that the malware’s language localization and distribution suggest a financial motivation, primarily targeting users in Argentina. Indications are that the malware was developed in a Chinese-speaking environment, as seen in debug strings written in simplified Chinese.

Distribution and Evolution of PromptSpy

PromptSpy is distributed through a dedicated website and has not been available on Google Play. It is considered an advanced iteration of a previously unknown malware called VNCSpy, with initial samples traced back to uploads from Hong Kong. The website “mgardownload[.]com” delivers a dropper that, upon installation, impersonates JPMorgan Chase, urging users to enable app installations from unknown sources.

According to ESET, the dropper contacts its server to fetch a configuration file, which includes a link to another APK presented as an update. However, the configuration server was found to be inaccessible during the investigation, leaving the specific download link undisclosed.

Implications and Future Outlook

This development highlights the increasing sophistication of malware, as threat actors incorporate AI to create adaptable and resilient threats. PromptSpy’s ability to avoid uninstallation through invisible overlays poses a unique challenge, with the only solution being rebooting the device into Safe Mode.

As ESET points out, PromptSpy exemplifies the evolving nature of Android malware, utilizing generative AI to navigate and manipulate on-screen elements across diverse devices and layouts. This capability marks a shift toward more dynamic and persistent cybersecurity threats.

The Hacker News Tags:accessibility services, AI security, Android malware, cybersecurity threats, ESET research, Gemini AI, malware persistence, mobile security, PromptSpy, VNCSpy

Post navigation

Previous Post: Adidas Probes Possible Third-Party Data Breach
Next Post: Android AI Malware Uses Google’s Gemini for New Threats

Related Posts

Microsoft Warns Misconfigured Email Routing Can Enable Internal Domain Phishing Microsoft Warns Misconfigured Email Routing Can Enable Internal Domain Phishing The Hacker News
OAuth Consent: The New Phishing Threat Bypassing MFA OAuth Consent: The New Phishing Threat Bypassing MFA The Hacker News
Google Resolves Critical Security Flaws in Gemini CLI Tools Google Resolves Critical Security Flaws in Gemini CLI Tools The Hacker News
Citrix Bleed 2 Flaw Enables Token Theft; SAP GUI Flaws Risk Sensitive Data Exposure Citrix Bleed 2 Flaw Enables Token Theft; SAP GUI Flaws Risk Sensitive Data Exposure The Hacker News
Why Organizations Are Turning to RPAM Why Organizations Are Turning to RPAM The Hacker News
Preinstalled Apps on Ulefone, Krüger&Matz Phones Let Any App Reset Device, Steal PIN Preinstalled Apps on Ulefone, Krüger&Matz Phones Let Any App Reset Device, Steal PIN The Hacker News

Categories

  • Cyber Security News
  • How To?
  • Security Week News
  • The Hacker News

Recent Posts

  • Armornet Launches AI Security for Software Supply Chains
  • Critical NGINX HTTP/2 Vulnerability Allows Remote Code Execution
  • SpyCloud Predicts 2027 Cyber Threats: AI and Trusted Access Risks
  • 2027 Cybersecurity Awards: New AI Categories Announced
  • Malicious Agent Framework Releases Threaten Credentials

Pages

  • About Us – Contact
  • Disclaimer
  • Privacy Policy
  • Terms and Rules

Archives

  • November 2026
  • October 2026
  • September 2026
  • August 2026
  • July 2026
  • June 2026
  • May 2026
  • April 2026
  • March 2026
  • February 2026
  • January 2026
  • December 2025
  • November 2025
  • October 2025
  • September 2025
  • August 2025
  • July 2025
  • June 2025
  • May 2025

Recent Posts

  • Armornet Launches AI Security for Software Supply Chains
  • Critical NGINX HTTP/2 Vulnerability Allows Remote Code Execution
  • SpyCloud Predicts 2027 Cyber Threats: AI and Trusted Access Risks
  • 2027 Cybersecurity Awards: New AI Categories Announced
  • Malicious Agent Framework Releases Threaten Credentials

Pages

  • About Us – Contact
  • Disclaimer
  • Privacy Policy
  • Terms and Rules

Categories

  • Cyber Security News
  • How To?
  • Security Week News
  • The Hacker News

Copyright © 2026 Cyber Web Spider Blog – News.

Powered by PressBook Masonry Dark