
Red Teaming (Model Score)
The DEFENDAI platform is constantly running hundreds of open source and proprietary security probes and scoring AI models. AI application builders can refer to these scores (with reasons) to help decide whether to use these models in their applications.
Customers can also submit their custom models with KPI's and the DEFENDAI platform will

Provide an initial score
Regularly monitor as more probes get added
Regularly test and report current KPI values