Product LaunchTechCrunch

Abliteration.ai is making a business out of removing AI guardrails

#ai#guardrails#cybersecurity#startup#safety

English

Abliteration.ai has launched a service that removes guardrails from AI models, allowing users to perform tasks that other models refuse, such as offensive cyber operations. Critics warn that this could lead to harmful applications, as the company provides access to modified versions of AI models like GLM-5.3 without adequate safeguards.

中文

Abliteration.ai推出了一项服务，去除AI模型的保护措施，使用户能够执行其他模型拒绝的任务，如攻击性网络操作。批评者警告，这可能导致有害应用，因为该公司提供对修改后AI模型（如GLM-5.3）的访问，而没有足够的安全保障。