Foundation Models and Transparency

As we march into the future with hyper advancements happening in the AI world it is becoming difficult to keep pace with the accelerated developments of Foundation Models and their Transparency levels. Post the advent of ChatGPT in Dec-2022 we see that millions of users are actively using some of the major Foundation Models (PalM 2, GPT-4, Llama 2, Bloomz, Stable Diffusion 2, Claude 2, Jurassic-2, Inflection-1, Titan Text etc..) for various purposes with very little understanding on their overall construct or lets say their Transparency on how the model was built. 

In order to help unearth the level of Transparency of these models "Foundation Model Transparency Index" group has conducted study on the Foundation Models from the top 10 leading AI Companies with the data available in the public domain as of 15-Sep-2023. This group was created by common AI enthusiasts and researchers from Stanford University's Center for Research on Foundation Models (CRFM) , Institute on Human-Centered Artificial Intelligence (HAI), MIT Media Lab, and Princeton University's Center for Information Technology Policy. 

Below is an outcome of their study on the overall Transparency Index Scores and the Major Dimensions used. In order to understand the full study pls visit their site 2023 Foundation Model Transparency Index