Telephone add-on devices to prevent special fraud, telephones and intercoms to prevent malicious door-to-door sales and forced purchases

Abstract

To provide a telephone add-on device, a telephone set, or an intercom that prevents special fraud by telephone and malicious door-to-door sales and forced buying by visitors. [Solution] This device is a telephone add-on device or telephone equipped with a generation AI processing unit that has been trained on telephone answering and specialized fraud, a switch unit that can block audio signals, and a mixing unit that superimposes the caller's voice and the audio signal generated by the generation AI processing unit, which avoids direct conversations with suspicious callers and prevents suspicious requests and contracts to prevent specialized telephone fraud. Also, an intercom equipped with a generation AI processing unit that has been trained on caller answering and malicious door-to-door sales and forced purchases, a switch unit that can block audio signals, and a mixing unit that superimposes the caller's voice and the audio signal generated by the generation AI processing unit, which avoids direct conversations with suspicious visitors and prevents malicious door-to-door sales and forced purchases to prevent suspicious requests and contracts.

Description

[Technical field]

[0001] This invention relates to a telephone add-on device for preventing special fraud, a telephone, and an intercom for preventing malicious door-to-door sales and forced buying. [Background technology]

[0002] Because telephones are highly convenient and can easily be used to call private phones, it has traditionally been difficult to prevent special telephone frauds against the elderly. In addition, there have been many cases of fraud in which unscrupulous renovation contractors visit the homes of elderly people and sign contracts for unreasonably high amounts, or where they try to force purchases of precious metals and other items at unreasonably low prices. However, these are difficult to distinguish from regular sales activities, making it difficult to take countermeasures. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent No. 7004983 Publication [Patent Document 2] JP 2020-112767 A [Patent Document 3] Patent No. 6670407 Publication Summary of the Invention [Problem to be solved by the invention]

[0004] Since the telephone numbers of fixed-line telephones connected to public lines have traditionally been published in telephone directories, elderly people are left defenseless against special frauds by malicious individuals over the phone, and the number of victims continues to grow. In addition, there have been many cases of malicious individuals visiting elderly people's homes and using clever words to commit fraudulent acts such as door-to-door sales and forced purchases. This device is intended to solve this problem. [Means for solving the problem]

[0005] This invention relates to a telephone add-on device for preventing specialized fraud, a telephone, and an intercom for preventing malicious door-to-door sales and forced buying. First, a telephone add-on device for preventing special fraud will be described. The telephone add-on device is inserted between the telephone body and the handset, and used by connecting with an audio signal cable. The present invention for the telephone add-on device is provided with a generation AI processing section (hereinafter referred to as the telephone response AI) that has learned telephone answering and special fraud methods, a switch section that can cut off the audio signal, and a mixing section that superimposes the caller's voice and the audio signal generated by the generation AI processing section, making it possible to transcribe the voice of the call into text and recognize the content of the conversation, determine whether there are any words, questions, requests, or other characteristics related to special fraud, generate messages to be conveyed to the other party depending on the situation, and speak the messages as audio. As a result, the telephone response AI will greet the caller in a polite manner by voice, listen to their name, the name of the organization they belong to, and the purpose of the call, and then make the message audible only to the person who picked up the receiver (hereinafter referred to as a family member). The AI ​​will first tell the caller's identity and the purpose of the call, and then ask the family member whether to transfer the call. However, if the AI ​​determines that the call is suspicious, it can give advice to the family member such as, "Shall I ask for the caller's contact number and tell the caller that I will call back after getting their contact number?". Even after the call has been transferred and a direct conversation with the other party has begun, if the AI ​​determines that the content of the call is suspected to be a specialized fraud, it can inform the family member so that only they can hear and give advice such as, "Shall I ask for the caller's contact number, giving a reason such as an emergency, and tell the caller that I will discuss it with my family and call back if necessary?". This will prevent malicious individuals from using the phone to induce family members to take actions that are detrimental to their interests, making it possible to prevent specialized fraud. Next, the telephone set for preventing special fraud is an integrated telephone set in which the above-mentioned telephone add-on device for preventing special fraud is built into the telephone set. This invention, which relates to an intercom that prevents malicious door-to-door sales and forced purchases, is equipped with a generation AI processing unit (hereinafter referred to as the visitor response AI) that has been trained in responding to visitors and malicious door-to-door sales and forced purchases, a switch unit that can block audio signals, and a mixing unit that superimposes the caller's voice and the audio signal generated by the generation AI processing unit, and is capable of transcribing the audio of the call into text and recognizing the content of the conversation, determining whether there are any characteristics such as words, questions, or requests related to malicious door-to-door sales and forced purchases, generating messages to be conveyed to the other party depending on the situation, and speaking these messages as audio. As a result, the visitor response AI will greet the visitor in a polite manner by voice, ask the visitor's name, the name of the organization they belong to, and the purpose of the visit, and then When a visitor arrives and presses the talk button on the intercom, the AI ​​will announce the caller's identity and purpose so that only the person (hereafter referred to as the family member) can hear, and then confirm with the family member whether to transfer the call. However, if the visitor response AI determines that the visitor is suspicious, it can give advice to the family member such as, "Shall we tell the caller that we're not interested and ask them to leave?". Even after the call has been transferred and a direct conversation has begun, if the visitor response AI determines that the content of the call is suspected to be malicious door-to-door sales or forced purchases, it can convey this so that only the family member can hear, and give advice such as, "Shall we tell the caller that we can't make a decision without consulting with our family and ask them to leave?", making it possible to prevent malicious door-to-door sales and forced purchases from happening in the first place. Effect of the Invention

[0006] As described above, by incorporating a telephone answering AI that has been trained in telephone answering and specialized fraud methods, a switch unit that can block audio signals, and a mixing unit that superimposes the caller's voice with the audio signal generated by the generation AI processing unit, it is possible to avoid calls between family members and suspicious people whose identities and purpose are unclear.In addition, even after the call is taken over and a direct conversation has begun between the caller and the family member, if the content of the conversation is recognized and judged and fraud is suspected, a warning can be issued to the family member and the call can be terminated, and the call can be ended on the family member's behalf, thereby preventing telephone fraud. In addition, by incorporating a visitor response AI that has been trained on how to respond to visitors and to malicious door-to-door sales and forced purchases, a switch section that can block audio signals, and a mixing section that superimposes the caller's voice and the audio signal generated by the generation AI processing section, it is possible to avoid conversations between malicious visitors and family members.Even after the call is taken over and a direct conversation between the visitor and family member begins, the content of the conversation is recognized and judged, and if malicious door-to-door sales or forced purchases are suspected, a warning is issued to the family member and the call is advised to be terminated, and a polite refusal is conveyed on the family member's behalf to the other party, thereby preventing malicious door-to-door sales and forced purchases. [Brief description of the drawings]

[0007] [Figure 1] 1 is an embodiment showing the main components of the present invention and the flow of signals and information. [Diagram 2] 13 is an embodiment of a generation AI processing unit. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0008] When a malicious person calls or visits a home and rings the intercom, elderly people who may be victims do not respond directly, but rather an AI is built in that can learn greetings, conversation content, and procedures and respond by voice, and the AI ​​is instructed to initially respond to the caller or visitor by voice. The greeting or response message generated by the AI ​​is superimposed as an audio signal by the mixing unit and transmitted to the caller or visitor as audio. In addition, to avoid direct conversation between malicious people and elderly people, the voice signal of the elderly is blocked by the switch unit. The AI ​​checks whether the caller has any suspicious characteristics, and if there are any suspicious characteristics, it does not connect the call and declines the suspicious visitor. Even after a direct call has begun, the system is equipped with AI that has been trained to detect specialized fraud, malicious door-to-door sales, and forced purchases.The AI ​​analyzes the conversation between the two parties, and if it determines that there is suspicion of specialized fraud, malicious door-to-door sales, or forced purchases, it will intervene in the conversation and block the voice signal of the malicious person using a switch, and then give appropriate advice to elderly people and others who may be at risk of becoming victims, and the AI ​​will take action on their behalf, such as hanging up the phone or refusing the call. EXAMPLES

[0009] An embodiment will be described below with reference to the attached drawings. Figure 1 shows the main components of the present invention and the flow of signals and information. 1 is a generation AI processing unit, and a specific embodiment thereof is shown in Figure 2 and will be described in the next section. Inside the generation AI processing unit, there are function blocks for voice recognition, caller speech analysis, receiver speech analysis, special fraud determination, telephone answering message generation, call intervention message generation, voice synthesis, and overall call control, and the connections and information flow between the function blocks are as shown in Figure 1. 2a and 2b are switch units that block the caller's voice signal according to a control signal from the generation AI processing unit as necessary. 3a and 3b are mixing units that superimpose the caller's voice signal and the voice signal generated in the generation AI processing unit to convey the message generated by the AI ​​to the caller as voice.

[0010] Figure 2 shows an example of the generation AI processing unit. 1 is an AI processor. Based on the AI ​​learning model, it processes the inference processing of the generation AI at high speed. 2 is an AI learning model. It is the data that is the basis for the processing such as recognition, inference, judgment, and generation performed by the generation AI. Depending on the application, it is made to learn the characteristics of telephone answering, visitor response, special fraud, malicious door-to-door sales, and forced purchase. 3 is a voice signal processor. It converts the caller's digital voice signal into digital voice data, converts the digital voice data generated by the generation AI into a digital voice signal, and performs other processing. 4 is a communication / input / output module. It converts digital voice signals to analog voice signals, inputs and outputs analog voice signals, and communicates via WiFi and mobile networks. Communication with the outside is mainly for the purpose of notifying a specified address by email of calls or visits from suspicious people, and incident information judged to be suspected of fraud, and for the purpose of updating the AI ​​learning model to improve it. 5, 6, and 7 are a CPU, memory, and secondary storage device, respectively, and function as a computer. The software for operating the generation AI processing unit is loaded from the secondary storage device into memory and executed, and the memory area used by the AI ​​processor is allocated. 8 is a computer bus that transmits data between the connected CPUs, memories, etc. 9 is a power supply that supplies power to each part in the device.

[0011] The embodiment shown in Fig. 1 has the same main components and can be applied to both a telephone add-on device for preventing special frauds and a telephone for preventing special frauds. Furthermore, based on the embodiment shown in Fig. 1, if the telephone is replaced with an intercom, the act of making a phone call is simply replaced with the act of visiting a house and ringing the intercom chime, and the AI ​​learning model used in the generation AI processing unit is replaced from one that has learned how to answer the phone and the methods of special fraud to one that has learned how to answer visitors and malicious door-to-door sales and forced purchases, it can be applied as an embodiment of an intercom that prevents malicious door-to-door sales and forced purchases. [Explanation of symbols]

[0012] Figure 1 1. Generative AI processing section 2 Switch section 3. Mixing Section

[0013] Figure 2 1 AI processor 2. AI Learning Model 3 Audio Signal Processor 4 Communication / I / O module 5 CPU 6. Memory 7 Secondary storage 8 Computer Bus 9 Power supply

Claims

1. A telephone add-on device that is inserted between a telephone and a handset and connected by an audio signal cable, and is equipped with a generation AI processing unit that has been trained on telephone answering and specialized fraud, a switch unit that can cut off audio signals, and a mixing unit that superimposes the caller's voice with the audio signal generated by the generation AI processing unit, and when a caller makes a call from outside and attempts to receive a call from the telephone, the device answers the call by voice on behalf of the recipient who receives the call, asking about the caller's identity and the purpose of the call, and if it determines that the call is suspicious, it will not transfer the call to the recipient, and if it does transfer the call, it recognizes and analyzes the content of the conversation between the caller and recipient, and if it determines that the content is suspicious, it will intervene in the call by warning the recipient or cutting off the call with the caller, thereby preventing specialized fraud.

2. A telephone set that incorporates a generation AI processing unit that has been trained in telephone answering and specialized fraud, a switch unit that can block audio signals, and a mixing unit that superimposes the caller's voice with the audio signal generated by the generation AI processing unit, and for callers who make an external call to the telephone set and attempt to speak to the caller, the telephone set answers the call by voice on behalf of the recipient who receives the call, asking about the caller's identity and the purpose of the call, and if it determines that the call is suspicious, it will not transfer the call to the recipient, and if it does transfer the call, it recognizes and analyzes the content of the conversation between the caller and recipient, and if it determines that the content is suspicious, it will intervene in the call by warning the recipient or cutting off the call with the caller, thereby preventing specialized fraud.

3. This intercom is equipped with a generation AI processing unit that has been trained on how to deal with visitors and malicious door-to-door sales and forced purchases, a switch unit that can cut off audio signals, and a mixing unit that superimposes the caller's voice and the audio signal generated by the generation AI processing unit, and responds to visitors by voice on behalf of the family member receiving the visit, asking about the caller's identity and purpose, and if it determines that the visit is suspicious, it will not transfer the call to the family member, and if it does transfer the call, it recognizes and analyzes the content of the conversation between the visitor and the family member, and if it determines that the content is suspected to be a special fraud, it will intervene in the call by warning the family member or cutting off the call with the visitor, thereby preventing malicious door-to-door sales and forced purchases.

Images