Method for communication between a client and an artificial intelligence agent, the method including location information

Abstract

The invention relates to a method for communication between a client and an artificial intelligence agent associated with at least one artificial intelligence model, the method comprising: - transmitting (201), by an intermediate device between the client and the artificial intelligence agent, client identification information, or transmitting, by the artificial intelligence agent or the intermediate device, artificial intelligence agent identification information; - verifying (202), by the client or the intermediate device, the artificial intelligence agent identification information, or verifying, by the artificial intelligence agent, the client identification information, according to a predefined verification rule.

Description

Description Title: Communication method between a client and an artificial intelligence agent including location information technical field

[0001] The present invention relates to the field of artificial intelligence and more specifically to the communication protocols between a user client and an artificial intelligence agent.

[0002] The invention aims in particular to enable an artificial intelligence agent to know a client's location information, in particular geographic location information, which transmits a request to it in order to apply access rules to the services provided by the agent based on this location information.

[0003] The invention also aims to enable a client wishing to communicate with an artificial intelligence agent hosted on a remote server to know the agent's location information in addition to the response provided by the agent to a client request. Previous technique

[0004] Artificial intelligence techniques, and in particular large language models (LLMs), are increasingly used in the execution of tasks such as producing a response text to an intention, diagnosing an incident, resolving that incident, resolving a malfunction, and optimizing a function.

[0005] In this context, communications assisted by an artificial intelligence agent, particularly generative artificial intelligence, are multiplying.

[0006] An artificial intelligence agent generates a response to a request from a client user independently of any information about the user or the client software.

[0007] In particular, there is no way for an artificial intelligence agent to filter its access to users based on certain criteria, especially geographical location criteria.

[0008] For example, in the context of a company with an internal generative AI agent and employees using that agent, there is no possibility that the AI ​​agent could condition its response on the employee's geographic location. Conversely, there is no possibility that the user's client software could be informed of the AI ​​agent's geographic location. However, for some companies, such as banks, this information is important for security reasons and due to industry regulations. Without a guarantee regarding a client's geographic location, a company might prefer to cancel communication with an AI agent rather than risk fraud.

[0009] The present invention aims to overcome the limitations of the prior art.

[0010] Summary of the invention

[0011] The purpose of this disclosure is to address some or all of the limitations of prior art solutions, including those described above, by proposing a solution that allows an artificial intelligence agent to receive location information from a client wishing to communicate with that agent. Similarly, the proposed solution also allows a client querying an artificial intelligence agent to receive the agent's location information in addition to the response to their query.

[0012] To this end, a communication method is proposed between a client and an artificial intelligence agent associated with at least one artificial intelligence model, the method comprising: A transmission, by the client or an intermediary device between the client and the artificial intelligence agent, of customer identification information, or a transmission, by the artificial intelligence agent or the intermediary device, of artificial intelligence agent identification information; A verification, by the client or the intermediary device, of the artificial intelligence agent's identification information, or a verification, by the artificial intelligence agent or the intermediary device, of the client's identification information, according to a predefined verification rule.

[0013] This process has the advantage of allowing the artificial intelligence agent to condition its response on information characterizing the client making the request, in particular geographic location information.

[0014] According to one aspect of the invention, the identification information includes at least one location information or one certification information, and the verification of the identification information involves at least the verification of the location information or the authentication of the certification information.

[0015] Thus, the artificial intelligence agent can condition its response to the customer based on their geographical location, or based on the result of customer authentication.

[0016] According to one aspect of the invention, the location information includes at least one piece of information from a source address or an access network identifier, and the certification information includes at least one certificate from the client, the artificial intelligence agent, or an access network.

[0017] This information allows us to characterize the customer's geographical location and to certify their identity.

[0018] According to a particular aspect of the invention, the method includes the artificial intelligence agent receiving a request from the client, the request including the client's identification information, and the artificial intelligence agent transmitting the content of the request to at least one artificial intelligence model if the client's identification information is valid according to said verification rule.

[0019] In this way, the artificial intelligence model is only queried if the customer identification is valid.

[0020] According to a particular aspect of the invention, the method includes a prohibition of access to the artificial intelligence model if the customer's identification information is not valid according to said rule, a transmission by the artificial intelligence agent to the customer, of a response from at least one artificial intelligence model to said request if the customer's identification information is valid according to said rule.

[0021] Thus, the customer is informed that they are not authorized to access the artificial intelligence model if their identification is not valid.

[0022] According to a particular aspect of the invention, the method comprises a transmission, by the client or the intermediate device, of a request to the artificial intelligence agent, a reception, by the client, of a response from at least one artificial intelligence model to said request including the identification information of the artificial intelligence agent.

[0023] Thus, the customer can receive identification information for the artificial intelligence agent, which allows them to associate with their response information characterizing the identity of the agent and, in particular, the agent's supplier.

[0024] According to a particular aspect of the invention, the method comprises receiving a request from the client through the intermediate device, adding the client's identification information to the request through the intermediate device, and transmitting the request through the intermediate device to the artificial intelligence agent.

[0025] In this way, an intermediary device is responsible for managing the client's identification information and adding it, if necessary, to the request sent to the artificial intelligence agent. This allows this functionality to be offloaded from the client to the intermediary device.

[0026] According to a particular aspect of the invention, the method includes an identification, by the intermediate device, of the access network to which the customer is connected, called the customer access network, a query, by the intermediate device, of the customer access network to retrieve an identifier of the customer access network, the customer identification information including at least the identifier of the customer access network.

[0027] The identifier of the access network to which the client connects allows the client's geographical location to be characterized.

[0028] According to a particular aspect of the invention, the method comprises receiving, by the intermediate device, customer access network certification information, and transmitting, by the intermediate device to the artificial intelligence agent, the customer access network certification information.

[0029] Communicating a certificate from the access network to which the client connects to the artificial intelligence agent allows the latter to authenticate the origin of the request.

[0030] According to a particular aspect of the invention, the method comprises receiving, by the intermediate device, a response from the artificial intelligence agent to the client's request, and transmitting, by the intermediate device, the response to the client.

[0031] According to a particular aspect of the invention, the method comprises a determination, by the intermediate device, of the location information of the artificial intelligence agent, and an addition of the location information of the artificial intelligence agent to the response to the customer.

[0032] The intermediary device retrieves the location information from the artificial intelligence agent and communicates it to the client. In this way, the functionality of adding this information to the response generated by the agent is managed by the intermediary device.

[0033] According to a particular aspect of the invention, the method includes receiving, by the intermediate device, the certification information of the artificial intelligence agent, and adding the certification information of the artificial intelligence agent to the response to the customer.

[0034] The intermediary device retrieves the certificate from the artificial intelligence agent and transmits it to the client. In this way, the functionality of adding this certificate to the response generated by the agent is handled by the intermediary device.

[0035] According to a particular aspect of the invention, the determination of the location information of the artificial intelligence agent includes an identification, by the intermediate device, of the access network on which the artificial intelligence agent is hosted, called the agent access network, and a query, by the intermediate device, of the agent access network to retrieve an identifier of the agent access network.

[0036] Identifying the access network on which the artificial intelligence agent is hosted thus makes it possible to characterize the agent's location.

[0037] According to a particular aspect of the invention, the method includes receiving, by the intermediate device, certification information from the agent access network, and adding the certification information from the agent access network to the response to the client.

[0038] The agent access network certificate is thus communicated to the client, who can authenticate the sender of the response if they wish.

[0039] The invention also relates to a communication device comprising a processor coupled to a communication device and a memory in which program instructions are stored for execution by the processor to carry out the steps of the communication process according to any one of the implementation modes of this disclosure.

[0040] The invention also relates to a computer program product comprising a set of program code instructions which, when executed by at least one processor, configure said at least one processor to implement the communication method according to any one of the implementation modes of this disclosure.

[0041] The invention also relates to a computer-readable recording medium on which is recorded a set of program code instructions which, when executed by at least one processor, configure said at least one processor to implement the communication method according to any one of the implementation modes of this disclosure. Brief description of the drawings

[0042] The invention will be better understood upon reading the following description, given by way of non-limiting example, and made with reference to the figures which represent: [Fig. 1]: a schematic representation of an artificial intelligence module according to a first embodiment, [Fig. 2a]: a diagram illustrating a communication system between a client and an artificial intelligence agent via an intermediary device, [Fig. 2b]: a diagram illustrating the general steps of a communication method between a client and an artificial intelligence agent via an intermediary device, [Fig. 3]: a diagram illustrating the steps of a method for registering a customer with an access network, [Fig. 4]: a diagram illustrating the steps of a method for identifying a customer using an artificial intelligence agent, [Fig. 5]: a diagram illustrating the steps of an example implementation of the communication method in Figure 2b, [Fig. 6]: a representation of a communication device capable of implementing one of the aforementioned methods.

[0043] In these figures, identical references from one figure to another designate identical or analogous elements. For clarity, the elements shown are not to scale unless otherwise indicated. Description of the implementation methods

[0044] More generally, it should be noted that the implementation and realization methods considered below are described as non-limiting examples, and that other variants are therefore conceivable.

[0045] Figure 1 schematically illustrates an example of an artificial intelligence module operating as an AI agent (AG).

[0046] Such an AG module is configured to receive a Req or "prompt" request from a client user and directed to an artificial intelligence (AI) model, for example, a generative model. The AG module includes a preprocessing module (PreProc) and a postprocessing module (PostProc). The PreProc and PostProc modules interact with a DTA database, an OTL tools module, and other MDL models. The MDL models are organized as functions that can be queried through an application programming interface (API). In one embodiment, the other MDL models are a subset of the OTL tools module.

[0047] The artificial intelligence model can be any type of known model, for example a generative model such as an LLM model (“large language model”).

[0048] Upon receiving the Req request, it is processed by the PreProc preprocessing module, which transforms the request content into data usable by the artificial intelligence model. For example, in the context of natural language processing (NLP) and multimedia processing, the PreProc preprocessing module transforms the request content into tokens. This "tokenization" operation is used to convert the request content, such as text, into an intermediate representation usable by the AI ​​model.

[0049] The PostProc post-processing module is configured to then transform data it receives from the artificial intelligence model into a response readable by the user client.

[0050] Figure 2a represents a simplified diagram of a communication system between a client CL and an artificial intelligence agent AG.

[0051] The CL client is, for example, a terminal such as a mobile communication device or a computer. The CL client is connected to a first network device R_CL or access network associated with a first telecommunications operator.

[0052] The artificial intelligence agent AG, for example, is implemented according to the architecture of Figure 1 to provide an artificial intelligence service to the client user CL.

[0053] The artificial intelligence agent AG is connected to a second network device R_AG or access network associated with a second telecommunications operator.

[0054] The first and second telecommunications operators may be different or the same. If they are the same, in a particular embodiment, the first network equipment R_CL and the second network equipment R_AG are identical.

[0055] The client user CL and the artificial intelligence agent AG communicate via an intermediate device PR or proxy, router or gateway which is connected to both network equipment R_CL, R_AG.

[0056] Figure 2b illustrates the main steps of a communication method between the client CL and the artificial intelligence agent AG.

[0057] The method may include a transmission step 201 by the client CL or the intermediate device PR, of identification information for the client CL. Alternatively or in combination, the method may include, in transmission step 201, a transmission, by the artificial intelligence agent AG or the intermediate device PR, of identification information for the artificial intelligence agent AG.

[0058] The method may further include a verification step 202, by the client CL or the intermediate device PR, of the identification information of the artificial intelligence agent AG. Alternatively or in combination, the method may include, at the verification step 202, a verification, by the artificial intelligence agent AG or the intermediate device PR, of the identification information of the client CL, according to a predefined verification rule.

[0059] According to one embodiment, the identification information includes location information for the client or artificial intelligence agent or the access network R_CL, R_AG to which they are connected.

[0060] According to one embodiment, the identification information includes certification information of the client or artificial intelligence agent or access network R_CL, R_AG to which they are connected.

[0061] Figure 3 illustrates the main steps of a method for registering a CL client with an R_CL access network.

[0062] At step 300, the software embedded on the CL client registers with the first telecommunications operator that provides the first access network R_CL. To do this, the CL client sends a registration request to the access network R_CL. The request includes the client's network connection credentials. These connection credentials are typically stored in lower protocol layers, such as the network layer or layers below the network layer.

[0063] In step 301, the R_CL access network checks if the client's network login credentials are valid. If the client's network login credentials are valid, the R_CL access network stores the client's source IP address and an R_CL access network identifier. Otherwise, the client registration is rejected.

[0064] Figure 4 illustrates the main steps of a method for identifying a customer CL with an artificial intelligence agent AG.

[0065] The method in Figure 4 is typically executed after the CL client has registered with the R_CL access network.

[0066] Client identification CL consists, for example, of retrieving a user token which is issued by the artificial intelligence agent AG.

[0067] Before a client can send a request to the AI ​​agent (hereinafter also referred to as the AG agent), a client must obtain access authorization in the form of a user token. This token represents, for example, a unique identifier for a session or interaction between the client and the AG agent.

[0068] Typically, it fulfills several functions, for example, client authentication via verification of the identity of the user making the request, session management which allows the interaction state to be maintained, enabling the AI ​​model to provide contextually relevant responses, or a rate limitation which can be used to track usage and apply limits to the number of requests a user can make.

[0069] User identity verification is done, for example, by opening an account associated with a username and password.

[0070] In step 400 of the method, the CL client transmits its connection credentials to the intermediate device PR which is connected to the first access network R_CL.

[0071] At step 401, the intermediate PR device sends a certificate indicating its identity to the client CL. This certificate is stored by the client. Step 401 is, for example, carried out via encrypted communication through the establishment of an encrypted tunnel between the client CL and the intermediate PR device, for example, a tunnel implementing the TLS (Transport Layer Security) protocol.

[0072] At step 402, the intermediate PR device transmits the CL client's login credentials to the AG artificial intelligence agent.

[0073] At step 403, the AG agent sends a certificate indicating its identity to the PR intermediate device. This certificate is stored by the intermediate device. Step 403 is, for example, carried out via encrypted communication through the establishment of an encrypted tunnel between the AG artificial intelligence agent and the PR intermediate device, for example, a tunnel implementing the TLS (Transport Layer Security) protocol.

[0074] At step 404, the AG agent verifies the CL client's login credentials. If the CL client's login credentials are valid, the AG artificial intelligence agent transmits an access authorization in the form of a user token to the PR intermediary device. The verification of login credentials includes, for example, verifying the user's identity or verifying that the user has sufficient funds in their account. The transmission of the user token is typically carried out via the HTTP protocol layer.

[0075] At step 405, the intermediate PR device transmits the user token and the agent certificate to the CL client. This step is, for example, carried out via the HTTP protocol layer.

[0076] The client identification method described in Figure 4 is preferably executed at least once before the first invocation of the AG agent by the CL client. It can be renewed to obtain a new access authorization (a new user token), for example, if the previous authorization is lost. It can also be renewed by the CL client to obtain the agent's certificate again. The certificate allows the CL client to identify the AG agent and, indirectly, the provider of the invoked artificial intelligence model, even though the client and the agent are not in direct communication due to the presence of the PR intermediary device.

[0077] Figure 5 illustrates an example of the implementation of the communication method between a client and an artificial intelligence agent via an intermediary device previously introduced in Figure 2b.

[0078] At step 500, the CL client connects to the PR intermediate device and sends a request, also called a "prompt", to the AG agent along with its user token obtained via the prior identification phase (Figure 4). This step is, for example, carried out via the HTTP protocol layer.

[0079] Optionally, the PR intermediary device sends a certificate indicating its identity to the CL client. This certificate is stored by the client. This step is performed, for example, via encrypted communication through the establishment of an encrypted tunnel between the CL client and the PR intermediary device, such as a tunnel implementing the TLS (Transport Layer Security) protocol.

[0080] In step 501, the PR intermediary device identifies the provider of the client's source IP address. This source IP address is obtained, for example, from the client's connection request (step 400) to the PR intermediary device. If the provider is a telecommunications operator unknown to the PR intermediary device, the PR identifies a web service implemented by that provider and deduces an identification information for the access network R_CL to which the client is connected.

[0081] In an optional 502 step, the client access network R_CL transmits a certificate to the intermediate device PR to authenticate the access network R_CL.

[0082] At step 503, the intermediate PR device requests a customer access network identifier (R_CL) associated with the source IP address from the network function implemented by the provider. The access network (R_CL) transmits this identifier to the intermediate PR device.

[0083] At step 504, the intermediate PR device adds the client's location information to the request. This information may include the location of the client's access network. For example, this location information is an identifier of the access network R_CL to which the client connected. This identifier may include a country identifier or, more generally, a geographic area identifier where the client is located. The client's location information is, for example, inferred from its source IP address.

[0084] At step 505 the intermediate PR device connects to the AG agent and transmits the client request along with the user token and the client's location information.

[0085] In an optional step 506, the PR intermediate device also transmits the certificate of the network that provided the customer access network identification.

[0086] In an optional step 507, the agent AG distributes a certificate to authenticate its identity. This certificate is stored by the intermediate device PR. Step 507 is, for example, carried out via encrypted communication through the establishment of an encrypted tunnel between the agent AG and the intermediate device PR, for example, a tunnel implementing the TLS (Transport Layer Security) protocol.

[0087] In step 508, the AG agent receives the information transmitted by the PR intermediary device. In step 509, it verifies whether the client's access network location information and / or the client's location information corresponds to an authorized geographic location. This verification involves, for example, comparing the location information to a predefined validity rule. For example, the validity rule specifies that users located in certain geographic areas are not authorized to send requests to the artificial intelligence model.

[0088] Step 509, for example, is executed by the AG agent's PreProc preprocessing module, which, based on the verification result, decides either to forward the client's request (step 510) to the AI ​​model or to reject the request by sending the client a response denying them access to the model (step 511). In a variant (step 512), the AG agent also checks whether the client's network access certificate is valid.

[0089] In one embodiment, steps 509, 511 and 512 are executed by the intermediate device PR and not by the agent AG.

[0090] The PreProc pre-processing module forwards the request to the AI ​​artificial intelligence model, for example, according to the mechanism described previously in support of Figure 1.

[0091] At step 513, agent AG transmits the response generated by the AI ​​model to the intermediate device PR.

[0092] At step 514, the intermediate device PR identifies the provider of the agent AG's IP address. If this provider is a different telecommunications operator than the client's and the intermediate device PR knows of a web service implemented by this provider, it queries this service, for example via the HTTPS protocol, to obtain identification information for the access network of the agent R_AG.

[0093] In an optional step 515, the access network of the agent R_AG transmits a certificate to the intermediate device to authenticate it. Step 515 is, for example, carried out via encrypted communication through the establishment of an encrypted tunnel between the access network of the agent R_AG and the intermediate device PR, for example, a tunnel implementing the TLS (Transport Layer Security) protocol.

[0094] At step 516, the intermediate device PR receives the response from the agent AG. It identifies the source IP address of the agent AG from this response and requests an access network identifier associated with this source IP address from the network function of the access network provider of the agent R_AG. The access network of the agent R_AG transmits this identifier to the intermediate device PR.

[0095] In step 517, the intermediate device PR forwards the response from agent AG to client CL. In one embodiment, in step 518, the intermediate device PR adds to the agent's response location information for agent R_AG's access network and / or location information for agent AG (for example, its source IP address). In another embodiment, in step 519, it also adds the authentication certificate of agent AG and / or the access network certificate of agent R_AG.

[0096] At step 520 the client receives the response and location information and, optionally, agent authentication information.

[0097] In the optional step 521, the CL client authenticates the AG agent. In a variant, this step 521 is performed by the PR intermediary device. This provides the client with the AG agent's geographic location information, which it can then use when evaluating the response or choose to query a different agent based in a different geographic area.

[0098] Figure 6 shows a diagram of an example DIS device configured to implement the processes described in Figures 2b to 5.

[0099] The DIS device includes a data processing module comprising a storage space, for example a MEM memory, a processing unit, equipped for example with a PROC microprocessor, and controlled by a computer program PGR whose instructions are configured to implement any of the processes as described in relation to Figures 2b to 5.

[0100] At initialization, the code instructions of the computer program PGR are, for example, loaded into MEM memory before being executed by the processing unit's processor. The processing unit's PROC microprocessor then implements, according to the instructions of the computer program PGR, the steps of any of the processes described in relation to Figures 2b to 5. To achieve this, in addition to the MEM memory and the PROC processor, the device includes COM communication means, enabling it to exchange messages with a server hosting an artificial intelligence agent. These communication means include, for example, a wired communication bus and can utilize communication methods such as an API, a Web API, and / or a communication protocol such as HTTP or SIP.

Claims

Demands

1. A method of communication between a client and an artificial intelligence agent associated with at least one artificial intelligence model, the method comprising: A transmission (201), by an intermediate device between the customer and the artificial intelligence agent, of customer identification information, or a transmission, by the artificial intelligence agent or the intermediate device, of artificial intelligence agent identification information; A verification (202), by the client or the intermediary device, of the artificial intelligence agent's identification information, or a verification, by the artificial intelligence agent, of the client's identification information, according to a predefined verification rule.

2. A method of communication between a client and an artificial intelligence agent according to claim 1, wherein the identification information includes at least one location information or one certification information, and the verification (202) of the identification information involves at least the verification of the location information or the authentication of the certification information.

3. A communication method according to claim 2, wherein the location information comprises at least one piece of information from a source address or an access network identifier, and the certification information comprises at least one certificate from the client, the artificial intelligence agent, or an access network.

4. A method of communication between a client and an artificial intelligence agent according to any one of the preceding claims, comprising: - Receipt (508) by the artificial intelligence agent of a request from the client, the request including the client's identification information, - A transmission (510) of the content of the request by the artificial intelligence agent to at least one artificial intelligence model if the client's identification information is valid according to said verification rule (509).

5. A communication method according to claim 4, comprising: - A ban on access (511) to the artificial intelligence model if the customer's identification information is not valid according to said rule, - a transmission (513) by the artificial intelligence agent to the customer, of a response from at least one artificial intelligence model to said request if the customer's identification information is valid according to said rule.

6. A method of communication between a client and an artificial intelligence agent according to any one of claims 1 to 3, comprising: - a transmission (500), by the client or the intermediary device, of a request to the artificial intelligence agent, - a receipt (520), by the client, of a response from at least one artificial intelligence model to said request including the identification information of the artificial intelligence agent.

7. A method of communication between a client and an artificial intelligence agent according to any one of claims 1 to 3, comprising: - A reception (500) by the intermediate device of a request originating from the client, - An addition (504) to the request, by the intermediary device, of the customer identification information, - A transmission (505) via the intermediate device of the request to the artificial intelligence agent,

8. A communication method according to claim 7, comprising: - An identification (501), by the intermediate device, of the access network to which the client is connected, called the client access network, - A query (503), by the intermediate device, of the customer access network to retrieve a customer access network identifier, the customer identification information including at least the customer access network identifier.

9. A communication method according to claim 8, comprising: - a reception (502), by the intermediate device, of customer access network certification information. - a transmission (506), via the intermediate device to the artificial intelligence agent, of the client access network certification information.

10. A communication method according to any one of claims 7 to 9, comprising: - A reception (516), by the intermediary device, of a response from the artificial intelligence agent to the client's request, - A transmission (517), via the intermediary device, of the response to the customer.

11. A communication method according to claim 10, comprising: - a determination (516), by the intermediate device, of the location information of the artificial intelligence agent, - an addition (518) of the artificial intelligence agent's location information to the response to the customer.

12. A communication method according to claim 11, comprising: - a reception (515), by the intermediate device, of the certification information of the artificial intelligence agent, - an addition (519) of the artificial intelligence agent's certification information to the response to the customer.

13. A communication method according to any one of claims 11 or 12, wherein the determination of the location information of the artificial intelligence agent comprises: - An identification (514), by the intermediate device, of the access network on which the artificial intelligence agent is hosted, called the agent access network, - A query (514), by the intermediate device, of the agent access network to retrieve an identifier of the agent access network.

14. A communication method according to claim 13, comprising: - the reception, by the intermediary device, of certification information from the agent access network, - an addition (519) of agent access network certification information to the response to the client.

15. Communication device comprising a processor (PROC) coupled to a communication device (COM) and a memory (MEM) in which program instructions (PGR) are stored for execution by the processor to carry out the steps of the communication method according to any one of the preceding claims.

Images