SaferAI is developing the auditing infrastructure for general-purpose AI systems (GPAIS), particularly large language models. To do so, SaferAI is pursuing both technical and governance work.
We're currently focused on demonstrating the existence of high-consequence risks among current models, and are working to develop standards for risk assessment of GPAIS.
We're extremely grateful to Founders Pledge who've been the main financial supporters of our work to tackle this incredibly pressing challenge.