US Patent Issued on Sept. 8 for "Systems and methods for enforcing a no-harm directive in artificial intelligence using constrained optimization and runtime safety shielding" (California Inventor)

ALEXANDRIA, Va., Sept. 8 -- United States Patent no. 12,730,879, issued on Sept. 8.

"Systems and methods for enforcing a no-harm directive in artificial intelligence using constrained optimization and runtime safety shielding" was invented by Anthony LaPine (Los Gatos, Calif.).

According to the abstract* released by the U.S. Patent & Trademark Office: "A system and method for enforcing a no-harm directive in artificial intelligence systems using a unified training-and-runtime control framework. A harm functional assigns quantitative harm scores to candidate actions in a given state or context, and a policy model is trained via constrained optimization to maximize a primary utility subject to harm constraints, including probability-of-harm...