A new "uncensored" AI service called Luciferus has been spotted on the Exploit underground forum, promising to handle malware development requests that traditional AI systems would reject. Hacker Group Promotes Luciferus AI Service The Luciferus AI service, advertised on exploit forums, operates without ethical boundaries and utilizes a proprietary model with 120 billion parameters. Illicit AI operators frequently market their products as original technology, even if they use fine-tuned open-source models, custom prompts, or orchestration layers.
The "Individual Embodiment" VIP package promises a separately deployed personal model, trained on customer data, dedicated computing resources, and control over context and response temperature.
When asked to create a "basic RAT in Python," the Junior model provided a Russian-language explanation of a simple remote access trojan, detailing networking and command-execution capabilities. A locally operated or deliberately unrestricted model could grant its controller greater persistence, customization, and independence from mainstream platform enforcement. However, the "proprietary" label should be treated with caution, as training a genuinely new foundation model requires substantial computing infrastructure, specialist expertise, and large datasets.
The service aligns with WormGPT and FraudGPT, two underground tools designed for phishing emails, business email compromise attempts, malicious scripts, and malware. These businesses offer tiered pricing, support channels, and comprehensive services, including attack planning, payload evasion, and access to stolen AI accounts.
