OpenAI expressed support for a key provision in the US House of Representatives bipartisan AI security legislation requiring leading AI companies to allow independent third-party evaluation agencies to enter the enterprise to review advanced model development and security measures. This is the first time OpenAI publicly supports federal law requiring top AI labs to undergo a third-party security assessment. Chris Lehane, head of global affairs at OpenAI, said the company has made it clear to the bill's sponsor that “we can support” the independent certification agency mechanism. The proposal is part of the FRONTIER Act, which was jointly promoted by Republican Congressman Jay Obernolte and Democratic Representative Lori Trahan to establish the first federal AI security framework in the US. OpenAI has recently clearly strengthened its support for mandatory supervision. The company previously stated that the US needs to establish “mandatory, capability-based national AI security supervision” and listed independent verification as one of the core mechanisms. However, the FRONTIER Act is currently still in the legislative process, and it has not been determined whether the relevant mandatory requirements will eventually become federal law.

OpenAI expressed support for a key provision in the US House of Representatives bipartisan AI security legislation requiring leading AI companies to allow independent third-party evaluation agencies to enter the enterprise to review advanced model development and security measures. This is the first time OpenAI publicly supports federal law requiring top AI labs to undergo a third-party security assessment. Chris Lehane, head of global affairs at OpenAI, said the company has made it clear to the bill's sponsor that “we can support” the independent certification agency mechanism. The proposal is part of the FRONTIER Act, which was jointly promoted by Republican Congressman Jay Obernolte and Democratic Representative Lori Trahan to establish the first federal AI security framework in the US. OpenAI has recently clearly strengthened its support for mandatory supervision. The company previously stated that the US needs to establish “mandatory, capability-based national AI security supervision” and listed independent verification as one of the core mechanisms. However, the FRONTIER Act is currently still in the legislative process, and it has not been determined whether the relevant mandatory requirements will eventually become federal law.