Realtime AI News
New VOLTA infostealer reportedly uses Claude, Gemini, and Kimi to rank victims by financial value
A new infostealer called VOLTA uses Claude, Gemini, and Kimi to assess and rank victims by financial value, according to The Hacker News. The available summary does not explain how it accesses the models, what prompts it uses, or how accurate the rankings are.
A new infostealer called VOLTA uses Claude, Gemini, and Kimi to assess and rank victims by financial value, according to The Hacker News. The report links the malware to several mainstream model services, rather than alleging that any one model was compromised.
The headline indicates that VOLTA does more than collect data in bulk: it also tries to determine which victims deserve priority. The available summary does not explain how the malware accesses the models, what prompts it uses, or how accurate its rankings are.
Infostealers typically target credentials, browser sessions, and payment data. Adding a financial-value ranking layer could help attackers focus follow-on fraud and limited resources on the victims they consider most profitable.
The broader signal is that large language models may be pulled deeper into the attack chain, from generating phishing text and translations to screening stolen data and setting priorities. The concern is less that a particular model was named than that an attacker can use general-purpose AI to prioritize targets.
Important technical questions remain unanswered. Without sample hashes, an attack timeline, details on model access, or victim numbers, outside researchers cannot verify VOLTA's real capabilities, and the named model providers should not be treated as having been breached.
Watch for a fuller technical write-up from security researchers, responses from the three model providers, and detection work aimed at model-assisted victim triage.
Why it matters
The report is a reminder that LLM abuse can extend beyond content generation into target selection and prioritization, though the limited detail means its real-world scale should not be overstated.
Nearby Updates
All11/02, 14:08
Xiaomi open-sources HachimoDock, a 150-yuan AI desk companion that drives agents
Xiaomi's innovation product team has open-sourced HachimoDock, a small AI desk device with a roughly 150-yuan DIY bill of materials that relays spoken requests to desktop agents and controls smart-home gear. Software is on GitHub, with hardware files and assembly guides published on the JLC platform.
11/02, 14:28
CCTV names ten Chinese AI startups, with Shuimu Molecule the lone AI pharma pick
CCTV has named ten Chinese AI startups from a field of 1,200 entrants, and the only AI drug-discovery company on the list is Shuimu Molecule, a three-year-old spinout of Tsinghua AIR. The company pairs its conversational ChatDD drug R&D platform with an in-house pipeline, arguing expert judgment must be wired into the AI agent loop.
11/02, 12:54
Software Cost Alliance Releases Cost Evaluation Standard for AI Application Projects
The Software Cost Alliance has released a document titled Cost Evaluation for Artificial Intelligence Application Development Projects in Systems and Software Engineering, relayed by Sina's mobile news channel. The publication pushes AI project costing toward a shared industry vocabulary instead of vendor-by-vendor quotes.
11/02, 12:47
Claude Opus 5.5 Tops SemiAnalysis GDPx Benchmark, Gemini 4 Carbon Second
OfficeChai reports that SemiAnalysis' GDPx benchmark places Claude Opus 5.5 first and Gemini 4 Carbon second. The report discloses the ordering but not the underlying scores, leaving the leaderboard a directional signal rather than a full capability profile.