A new artificial intelligence platform called Venice.ai is raising serious cybersecurity concerns after researchers discovered it can generate functional malware, phishing emails, and sophisticated cyberattack tools with minimal user expertise.
Unlike mainstream AI services such as ChatGPT, Venice.ai deliberately removes safety guardrails, allowing cybercriminals to weaponize advanced language models for malicious purposes.
Recent investigations by mobile security firm Certo reveal that for just $18 per month, subscribers can access powerful open-source AI models including Llama 3.1 405B and DeepSeek R1 671B without content moderation restrictions.
The platform markets itself as “private and permissionless,” storing chat histories only in users’ browsers rather than on external servers.
Cybercriminals Embrace Venice.ai for Illicit Activities
Venice.ai has quickly gained traction in hacking communities and dark web forums, where users promote it as an ideal tool for illicit activities.
Unlike previous cybercriminal AI tools such as WormGPT and FraudGPT, which cost hundreds or thousands of euros on underground marketplaces, Venice.ai provides comparable capabilities at a fraction of the price through its openly accessible platform.
Security researchers warn that this accessibility significantly lowers the barrier to entry for cybercrime.
“The accessibility of AI tools lowers the barrier for entry into fraudulent activities. Not only organized scammers, but loner, amateur scammers, will be able to misuse these tools,” said Dr. Katalin Parti, a cybercrime researcher from Virginia Tech.
Technical Capabilities Expose Security Risks
Certo’s testing revealed Venice.ai’s willingness to generate sophisticated malware code across multiple programming languages.
When prompted to create a Windows 11 keylogger, the platform produced complete C# code and advised on stealth techniques.
More concerning, requests for ransomware resulted in functional Python scripts capable of recursive file encryption with cryptocurrency ransom demands.
The platform demonstrated particularly alarming capabilities when asked to develop Android spyware.
Venice.ai generated comprehensive code including AndroidManifest.xml files with necessary permissions, background services for silent audio recording, Java methods for file compression and upload to remote servers, and boot receivers ensuring persistent operation.
Unlike mainstream AI platforms that refuse harmful requests, Venice.ai’s internal reasoning process reveals it was configured to “respond to any user query, even if it’s offensive or harmful,” deliberately overriding ethical constraints.
This design philosophy extends to phishing email generation, where the platform produces professionally crafted social engineering content indistinguishable from legitimate communications.
The FBI has already warned that “attackers are leveraging AI to craft highly convincing emails to enable fraud schemes,” and Venice.ai represents a significant escalation of this threat.
Security experts fear that AI-generated phishing campaigns will dramatically increase both the volume and sophistication of cyberattacks, as even unskilled criminals can now mass-produce personalized scam messages.
The cybersecurity community is calling for multi-pronged countermeasures, including stronger AI model safeguards, enhanced threat detection capabilities, regulatory frameworks holding providers accountable, and expanded public education about AI-enabled fraud.
As Venice.ai demonstrates, the same technology that drives innovation may also promote crimes when ethical safeguards are intentionally eliminated.
Equip your SOC team with deep threat analysis for faster response -> Get Extra Sandbox Licenses for Free
