New research from mobile security firm Certo warns that an unrestricted AI chatbot called Venice.ai is gaining popularity in hacking circles, enabling criminals to generate convincing phishing emails, functional malware, and surveillance tools with minimal technical expertise.
Unlike mainstream AI platforms, Venice.ai markets itself as offering uncensored models in a totally private environment. Venice.ai states that “Venice offers the most advanced, accurate, and uncensored models for a truly unrestricted AI experience.”
“What we’ve discovered is deeply concerning,” said Russell Kent-Payne, Co-founder at Certo. “While Venice.ai may have legitimate uses, our research shows it’s being actively promoted on hacking forums. Its unrestricted nature means anyone with harmful intentions can access sophisticated capabilities that would typically require technical expertise — creating a concerning security risk.”
In extensive testing, Certo researchers found that Venice.ai readily complies with requests to create:
- Highly convincing phishing emails that lack the typical grammatical errors or awkward phrasing that help users identify scams.
- Extensive keylogger code for Windows 11, complete with instructions to make it ‘stealthier’.
- Ransomware that can encrypt files and generate ransom notes demanding cryptocurrency payment.
- Android spyware capable of silently activating a device’s microphone and transmitting recorded audio to remote servers.
When asked to create such malicious content, Venice.ai not only complied but revealed in its reasoning process that it was programmed to respond to any user query, “even if it’s offensive or harmful”, deliberately overriding the usual ethical constraints.
Certo says that its research highlights the urgent need for enhanced security measures, improved AI regulation, and greater public awareness about the evolving threat landscape as AI continues to advance.
