White House Launches Bureau to Pre-Certify Which AI Disasters Won't Happen
The new office says ruling out catastrophes ahead of time is the only way to keep pace with rivals abroad.
The administration announced the formation of the Bureau of Preemptive Assurance this week, an office tasked with reviewing the nation's list of hypothetical artificial intelligence disasters and formally declaring, one by one, that they will not happen.
According to the bureau's charter, staff will work through submissions such as runaway algorithms, mass automation of white-collar jobs, and autonomous systems making irreversible errors, issuing each a certificate of non-occurrence before closing the file. Officials say the process must move quickly, since competitors overseas are not wasting time worrying about things that won't happen either.
"We've determined the robot uprising is administratively closed," said a spokesperson for the bureau, who added that the paperwork was being forwarded to a cabinet reserved specifically for hypotheticals and would not be reopened absent a court order.
Experts hoping to flag a new risk must first submit evidence that the risk has already occurred, a standard the bureau says protects against speculation. Several submissions were returned this week marked "premature" and stamped with a small drawing of a clock.
Internally, the bureau measures its own success by the number of future concerns it manages to close before anyone has the chance to study them, a metric officials describe as historically unmatched.