US 12,476,822 B2
Differentiation between human-generated and AI-generated text
Maxwell Robert Murialdo, San Francisco, CA (US)
Assigned to LAWRENCE LIVERMORE NATIONAL SECURITY, LLC, Livermore, CA (US)
Filed by Lawrence Livermore National Security, LLC, Livermore, CA (US)
Filed on Jun. 1, 2023, as Appl. No. 18/327,453.
Prior Publication US 2024/0406003 A1, Dec. 5, 2024
Int. Cl. H04L 9/32 (2006.01); H04L 9/08 (2006.01)
CPC H04L 9/3247 (2013.01) [H04L 9/0877 (2013.01)] 31 Claims
OG exemplary drawing
 
1. A method comprising:
generating a plurality of first cryptographic digital signatures associated with a set of typed characters, the set of typed characters having been input by use of a physical keyboard that is part of or in direct communication with a first processing device, wherein generating the plurality of first cryptographic digital signatures includes, for each of a plurality of keystrokes on the physical keyboard:
detecting a motion associated with a pressing of a particular key on the physical keyboard, wherein the particular key is associated with a particular character, said detecting including receiving first spatially dependent sensory data resulting from the plurality of characters being typed on the physical keyboard, by using a distributed array of sensor devices and verifying that the typed characters have been generated by a physical typing of keys on the physical keyboard, based on the first spatially dependent sensory data;
generating a first cryptographic digital signature associated with the particular character in response to detecting the motion associated with the pressing of the particular key; and
providing a typed character and an associated one of the first cryptographic digital signatures to a first verification entity located within the first processing device, for verification of a physical keystroke;
accessing a spatial mapping associated with a physical keyboard, the spatial mapping being based on second spatially-dependent sensory data associated with the physical keyboard, the second spatially-dependent sensory data being dependent on a position and orientation of the physical keyboard; and
wherein verifying that the typed characters have been generated by a physical typing of keys on the physical keyboard is based on a comparison of the first spatially-dependent sensory data and the spatial mapping.