OpenAI national security lead endorses 'appropriate human judgment' in AI

"Sasha Baker, head of National Security Policy at OpenAI, stated that there is a workforce transformation needed to educate analysts and service members to apply 'appropriate human judgment' in defense operations, especially as AI systems become more integrated."

"Baker highlighted that the consequences of incorrect AI-driven decisions in defense operations are much greater depending on the use case, underscoring the critical need for human oversight."

"She noted that AI transformation impacts the entire spectrum of national security work, from improving efficiency in paperwork to potentially revolutionizing the targeting cycle in military operations."

"Baker reiterated OpenAI's commitment to safety in AI deployment, mentioning that no large language model, including ChatGPT, is foolproof and emphasizing the importance of the Trusted Access program for ensuring safety."

Incorporating advanced AI capabilities into defense operations requires workforce transformation and education for analysts and service members to apply appropriate human judgment. The consequences of AI-driven decisions can be significant, especially in military contexts. OpenAI's head of national security policy emphasized the need for safety in AI deployment, noting that no large language model is completely foolproof. OpenAI has established a Trusted Access program to enhance safety in AI applications, particularly in sensitive areas like national security.

Read at Nextgov.com