Protecting private information during large language module interactions
Systems and methods for protecting private information during large language module (LLM) interactions. In one embodiment, an LLM gateway may route user prompts through a client server that possesses its own access credentials to an LLM server. The access credentials of the client server may be separate from and unassociated with the user. The client server may access the LLM using its own access credentials, submit the user's prompt to the LLM server, and return the results to the user. Routing user prompts through a client server that possesses its own LLM access credentials may help to ensure that user prompts cannot be traced back to the user or to personal information that a user disclosed in order to gain access to an LLM server.
1 . A computer-implemented method for protecting private information during large language module model (LLM) interactions, at least a portion of the method being performed by a computing device comprising one or more processors, the method comprising:
receiving a ranked listing of LLM providers based on levels of user privacy offered;
receiving a prompt from a user;
analyzing the prompt to determine a level of sensitivity associated with the prompt, wherein the level of sensitivity is determined based on an amount and/or degree of private information contained within the prompt;
selecting an LLM server to generate a response to the prompt based on the level of sensitivity associated with the prompt and a level of user privacy offered by the LLM provider associated with the LLM server;
routing the prompt through a client server that possesses access credentials to the LLM server, wherein the access credentials of the client server are unassociated with the user and prevent the LLM provider from associating the prompt submission with the user's identity via account credentials;
accessing, by the client server, the LLM server through the access credentials;
submitting, by the client server, the prompt to the LLM server; and
returning results from the LLM server to the user.
2 . The method of claim 1 , wherein the access credentials include a username and password.
3 . The method of claim 1 , wherein obtaining access credentials requires disclosure of personal information.
4 . The method of claim 3 , wherein the personal information includes a name and a birth date.
5 . The method of claim 1 , wherein geographical locations of the user and the client server are different.
6 . The method of claim 1 , further comprising:
receiving a second prompt from the user;
routing the second prompt through a second client server that possesses second access credentials to the LLM server, wherein geographical locations of the client server and the second client server are different;
accessing, by the second client server, the LLM server through the second access credentials;
submitting, by the second client server, the second prompt to the LLM server; and
returning results from the LLM server to the user.
7 . The method of claim 1 , wherein the private information includes personal information about the user or information that is not publicly known.
8 . The method of claim 1 , wherein the prompt is determined to have a low level of sensitivity and the LLM server is an open source LLM provider.
9 . The method of claim 1 , wherein the prompt is determined to have a high level of sensitivity and the LLM server is a private LLM provider.