Josephine Teo, the Minister of Communications and Information in Singapore, has unveiled AI Verify – Project Moonshot beta version, a suite of artificial intelligence testing tools designed to assess security related to the use of Large Language Models (LLM), an open-source platform capable of simulating attacks, benchmarking standards, and conducting foundational testing to address AI risks.
Developed in collaboration with partners such as DataRobot, IBM, Singtel, and Temasek, Project Moonshot aims to establish a global testing standard for leading AI organizations like the AI Verify Foundation (AIVF) and MLCommons. These organizations have signed a Memorandum of Intent (MOI) to collaborate in creating a shared security standard.
Key capabilities of Project Moonshot include performance and security evaluation of models using Hugging Face standards and MLCommons testing suite, assessing and scoring AI models to enhance confidence in their real-world deployment, and conducting mass prompt injection for Red Teaming security assessments.
Developers can now access the beta version of Project Moonshot on GitHub.
Source: PR Newswire, AI Verify Foundation
TLDR: Josephine Teo introduces AI Verify – Project Moonshot beta version, a tool for evaluating security risks associated with Large Language Models, developed in partnership with industry leaders to establish a global AI testing standard.
Leave a Comment