President Joe Biden is anticipated to sign a memorandum on Thursday outlining how intelligence agencies and national security agencies, including the Pentagon, should use and implement guardrails around AI. Reports The New York Times.
The order urges people to be told about AI tools that might be used as targeting weapons and prohibits AI from making decisions about granting asylum, persecuting an individual based on their ethnicity or religion, or classifying an individual as to have a “known terrorist” hit. without human verification.
Additionally, the memorandum calls for intelligence agencies to start protecting work on AI and AI chips from espionage or theft by foreign adversaries. It also authorizes the recently formed AI Safety Institute to assist vet AI tools before they’re released to make sure they can’t support terrorist groups or hostile nations.
As the New York Times notes, it's unclear how effective the order will ultimately be, since a lot of the deadlines it sets will expire after Biden leaves office.
