THE WHITE HOUSE SILENTLY DEMARCATED THE OPEN SOURCE MODEL WITHOUT THE ADVANCED AI TESTING FRAMEWORK

ON 5 AUGUST, ACCORDING TO AXIOS, THE WHITE HOUSE EXCLUDED THE OPEN SOURCE MODEL FROM THE ADVANCED AI CAPABILITY TEST FRAMEWORK. ACCORDING TO SEVERAL SOURCES OF INFORMATION ABOUT WHITE HOUSE MEETINGS, THE AI FRAMEWORK CONSIDERED ON TUESDAY DEFINES THE “COVERED FRONTIER MODEL” AS A CLOSED-SOURCE MODEL WITH STATE-OF-THE-ART CAPABILITIES AND NATIONAL SECURITY RISKS. THERE IS NO CLEAR DEFINITION OF WHAT CONSTITUTES “MOST ADVANCED” OR “NATIONAL SECURITY RISK”. THE WHITE HOUSE HELD A STAFF-LEVEL MEETING WITH INDUSTRY ON TUESDAY TO REVIEW THE FRAMEWORK, AND UNINVITED COMPANIES REMAIN UNAWARE OF ITS CONTENT. IT IS NOT CLEAR WHICH “CREDIBLE PARTNERS” WILL RECEIVE EARLY ACCESS TO ADVANCED MODELS UNDER THE FRAMEWORK, INCLUDING THE ELIGIBILITY OF FOREIGN GOVERNMENTS. ACCORDING TO SOURCES, COMPANIES WERE ENCOURAGED TO SUBMIT TO THE GOVERNMENT, TO THE EXTENT POSSIBLE, MODELS OF NEAR-PUBLICITY, RATHER THAN EARLY VERSIONS, AT THE TUESDAY MEETING. ACCORDING TO SOURCES, MANY COMPANIES WITH LESS ADVANCED SYSTEMS MAY BE EXCLUDED FROM THE FRAMEWORK。