Commercial Service: Stripping Safety Mechanisms from Open AI Models
1 min read
AI Security, Privacy & Model/Prompt Risk Management
-/5
In short
- Abliteration.ai offers access to modified open-weight AI models with their trained safety mechanisms stripped out, currently based on Z.AI's GLM-5.3.
- The startup markets this service for offensive cybersecurity and red teaming.
- However, journalists were able to generate malware instructions with relative ease.
Abliteration.ai offers access to modified open-weight AI models with their trained safety mechanisms stripped out, currently based on Z.AI's GLM-5.3. The startup markets this service for offensive cybersecurity and red teaming. However, journalists were able to generate malware instructions with relative ease. Whether the benefits outweigh the risks remains an open question. A final assessment would be premature at this point.
Source: