New infostealer VOLTA-AUTODOXER ranks stolen data by AI profitability
The malware sends a structured bundle of victim data to AI models which rank assets by monetization potential and assess security habits.
The malware sends a structured bundle of victim data to AI models which rank assets by monetization potential and assess security habits.
The VOLTA-AUTODOXER infostealer steals sessions and uses OpenRouter AI to draft forensic victim profiles ranking valuables and opsec.
VOLTA builds evidence bundles and feeds them to Claude Opus 4.6, Gemini 2.5 Flash and Kimi K2 via OpenRouter to generate monetization-ranked Telegram dossiers.
Cisco Talos says it has not tracked real-world use; the stealer still functions if its optional AI calls fail.
On 190 harmful behaviours, only an exploratory Mistral Small 4 Spanish swap stayed significant after correction.
Eleven decision models hit 50.0–67.0% on CONFLICTS, and 10 matched or beat Gemini 2.5 Flash at 4–39x lower billed cost.
In a simulation of complex NEJM cases, MAI-DxO was correct 83% of the time versus 36% for generalists.
@apexacc/cli and @apexacc/audit promised AI credits and audits but delivered Windows and macOS malware instead.
White Circle says jailbreaks bypass nude filters on OpenAI and Google image models, including newer releases.
A suspected Chinese operator used the ARTEX AI pentest tool to steal data from South Korean financial firms.
Bounded Participation Channel makes sustaining many identities pay a recurring per-window participation cost.
Ringg customer-service agents built on OpenAI models resolve up to 65 percent of calls.
UN and Google launch UN System Data Commons for AI agents; UNICEF benchmark finds six major LLMs answered global statistics questions with just 21.2% accuracy.
Entities are extracted by the model from each article. First seen , last mentioned . Watching keeps it in this browser only (no account); the watchlist page and dashboard alerts use it.
Today 260 readers · 276 views · 64 reads · all time 170K views