AI‑Powered ‘Volta‑Autodoxer’ Malware Harvests Victims’ Data at Scale

A new strain of infostealer malware dubbed Volta-Autodoxer has been identified by security researchers, and it leverages artificial‑intelligence tools to automate the extraction of a wide array of personal information from infected devices.

The code is designed to locate and exfiltrate everything from usernames and passwords to credit‑card numbers, Discord session tokens and even receipts generated by food‑delivery services such as UberEats. By scanning browsers, saved files and application caches, the program builds a comprehensive profile of each victim without requiring manual configuration.

What sets Volta‑Autodoxer apart is its integration of AI‑driven parsing engines. Instead of relying on static patterns, the malware invokes language‑model APIs to interpret unstructured data, recognize sensitive fields and decide which pieces are worth sending back to the command‑and‑control server. This dynamic approach allows it to adapt to new apps and services faster than traditional infostealers.

The emergence of AI‑enhanced threats reflects a broader trend in cybercrime, where attackers repurpose tools originally intended for legitimate automation and content generation. Over the past few years, several high‑profile campaigns have incorporated machine‑learning classifiers to evade detection, and Volta‑Autodoxer appears to be the latest iteration that pushes the capability into the data‑collection phase itself.

Industry analysts warn that the combination of automated credential harvesting and the ability to parse a victim's digital receipts could fuel more targeted fraud, credential stuffing and social‑engineering attacks. Security firms are urging users to keep software up to date, employ multi‑factor authentication wherever possible and monitor financial statements for unexpected activity. As AI becomes more accessible, experts expect similar techniques to appear in future malware families, prompting a need for defensive solutions that can recognise not just known signatures but also the behavioural hallmarks of AI‑driven exploitation.