US 12,277,246 B1
Derived permissions for stored outputs from artificial intelligence agents
John Manton, Cumming, GA (US); and Spencer Reagan, Marietta, GA (US)
Assigned to Airia LLC, Alpharetta, GA (US)
Filed by Airia LLC, Alpharetta, GA (US)
Filed on Nov. 29, 2024, as Appl. No. 18/964,322.
Claims priority of provisional application 63/658,434, filed on Jun. 10, 2024.
Int. Cl. G06F 21/62 (2013.01)
CPC G06F 21/6218 (2013.01) 20 Claims
OG exemplary drawing
 
1. A method for deriving permissions for stored outputs from artificial intelligence (AI) agents, comprising:
receiving a first input associated with an AI agent;
identifying multiple data chunks by causing comparison of a vectorized portion of the first input to embeddings in at least one vector database, the multiple data chunks being associated with different access permissions;
deriving a permission for association with a first output of the AI agent, resulting in a derived permission that comprises a least permissive intersection of the different access permissions of the multiple data chunks;
causing storage of the first output with metadata, wherein the metadata is associated with the derived permission, and wherein the first output and metadata are stored in a first vector database of the at least one vector database;
receiving a second input associated with the AI agent from a user device, the second input being associated with a user having a user permission;
in an instance in which the first output is responsive to the second input, causing a comparison of the user permission to the derived permission of the first output; based on the comparison, generating a second output based on the first output; and causing display, on the user device, of at least a portion of the second output.