Microsoft Corp.,\u00a0Alphabet Inc.\u2018s Google, and Elon Musk\u2019s xAI have pledged to grant\u00a0the U.S. government early access\u00a0to their latest artificial intelligence models to facilitate preliminary national security risk assessments.<\/p>\n
The Center for AI Standards and\u00a0Innovation (CAISI) under the Department of Commerce announced the agreement on Tuesday. The deal would allow the center to examine and research the AI models to determine their capabilities and potential security risks before their public deployment.<\/p>\n
The body has already conducted over 40 evaluations, including on advanced models that are not yet publicly available. Developers often submit\u00a0versions of their models with\u00a0safety guardrails removed, enabling the center to thoroughly assess national security risks.<\/p>\n
Don’t Miss:<\/p>\n
Chris Fall, the Director of CAISI, underscored the necessity of independent and rigorous measurement science in comprehending the impact of frontier AI on national security.<\/p>\n
This agreement builds on the 2024 deals with OpenAI and Anthropic, which were established during\u00a0the Biden administration when CAISI\u00a0was known as the U.S.\u00a0Artificial Intelligence Safety Institute.<\/p>\n
This agreement comes in the wake of recent national security concerns related to AI. Last week, the Pentagon\u2019s tech chief, Emil Michael, highlighted the potential security risks of Anthropic\u2019s Claude models. He also noted the advanced cyber capabilities of Mythos, another AI model, which could potentially identify and rectify cyber vulnerabilities.<\/p>\n