Dementia prevention system

Abstract

To provide a dementia prevention system which can be used for preventing a user from suffering from dementia or suppressing the progress of dementia.SOLUTION: The system includes a server (11) provided with a AI operator unit (111) for having a conversation by artificial intelligence, a terminal device (12) held by an object person (2) who is to be prevented from becoming demented or whose progress is to be suppressed, and a telecommunication network (13) for connecting the server (11) and the terminal device (12), wherein the AI operator unit (111) periodically makes a telephone call to the terminal device (12) of the object person (2) or receives a telephone call from the terminal device (12) of the object person (2) to have a conversation with the object person (2).SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a dementia prevention system that is preferably used to prevent the onset of dementia or suppress the progression of dementia. [Background technology]

[0002] Dementia is a condition characterized by the death or impaired function of brain cells due to various causes, resulting in various disorders that disrupt daily life and continue for approximately six months or more. The most common causes of dementia are "degenerative diseases," in which brain nerve cells slowly die. Examples of these "degenerative diseases" include Alzheimer's disease, frontotemporal dementia, and Lewy body disease. The second most common is vascular dementia, in which cerebral infarction, cerebral hemorrhage, or cerebral arteriosclerosis deprives nerve cells of nutrients and oxygen, resulting in the death of affected nerve cells and the breakdown of neural networks. Symptoms directly resulting from brain cell destruction are called core symptoms, including memory loss, disorientation, impaired comprehension and judgment, and impaired executive function. These core symptoms cause dementia patients to lose their ability to correctly perceive the reality of what is happening around them.

[0003] A dementia testing method based on a question-and-answer system using an AI call service on a server is known (see Patent Document 1). This dementia testing method involves providing guidance to a user device connected to the AI ​​call, conducting a preliminary practice test (first test) to improve the subject's understanding of the test procedure, and conducting a final test (second test) to determine the subject's presence or absence of dementia and the severity of their dementia symptoms. In the first test, the user device is requested to respond to a first topic during a first period, and a numerical value (comprehension value) representing the subject's understanding of the test procedure is calculated based on the responses. In the second test, the user device is requested to respond to a second topic different from the first topic during a second period longer than the first period, and at least one word corresponding to the second topic is extracted based on the responses. A language fluency score is calculated based on the extracted word corresponding to the second topic. Based on these comprehension scores and language fluency scores, patients who need to visit a testing center for subsequent testing are filtered out. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Special Publication No. 2024-502687 Summary of the Invention [Problem to be solved by the invention]

[0005] According to the conventional dementia testing methods described above, although it is possible to conduct a dementia test without the need for the subject to visit a testing center in person, there is a problem in that it is not possible to prevent dementia from occurring or to suppress the progression of dementia.

[0006] The problem to be solved by the present invention is to provide a dementia prevention system that can be used to prevent dementia or suppress the progression of dementia. [Means for solving the problem]

[0007] The present invention includes a server equipped with an AI operator unit that conducts conversations using artificial intelligence, a terminal device held by a subject whose onset of dementia should be prevented or whose progression of dementia should be suppressed, and a telecommunications network that connects the server and the terminal device, and the AI ​​operator unit solves the above problem by periodically calling the subject's terminal device or receiving calls from the subject's terminal device and conversing with the subject. [Effects of the Invention]

[0008] According to the present invention, the subject will have regular conversations, which will increase brain activity, relieve stress, and the trusting relationship with the caller will stabilize the mind, thereby preventing the onset of dementia and slowing the progression of dementia. [Brief explanation of the drawings]

[0009] [Figure 1] 1 is a block diagram showing an embodiment of a dementia prevention system according to the present invention. [Figure 2] 10 is a flowchart showing an embodiment of information processing executed by a server of the dementia prevention system according to the present invention. [Figure 3] 1 is a time chart showing an example of a conversation carried out in the dementia prevention system according to the present invention. [Figure 4] 10 is a time chart showing another example of a conversation carried out in the dementia prevention system according to the present invention. [Figure 5] 10 is a time chart showing yet another example of a conversation carried out in the dementia prevention system according to the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0010] System Configuration Hereinafter, with reference to the drawings, an example of an embodiment of the present invention will be described. FIG. 1 is a block diagram showing one embodiment of a dementia prevention system according to the present invention. The dementia prevention system 1 of this embodiment includes a server 11, one or more terminal devices 12, a telecommunications network 13 connecting the server 11 and the terminal devices 12, and a monitor 114, and provides the services described below to subjects who should be prevented from developing dementia, or subjects who already have dementia but whose progression should be slowed. Hereinafter, subjects who should be prevented from developing dementia and subjects who already have dementia but whose progression should be slowed will be collectively referred to as subjects 2.

[0011] The server 11 of this embodiment is configured as a computer including a CPU, ROM, and RAM, with predetermined software installed in the ROM, and is installed, for example, in a call center, etc. The call center referred to here is also called a customer center, customer service center, contact center, answering center, support center, etc., and refers to an office or department that specializes in handling complaints and inquiries from customers by telephone, facsimile, or software with a calling function, and refers to a large facility with a relatively large number of telephone lines and operators, such as the inquiry desk of a major company.

[0012] The server 11 includes an AI operator unit 111, an intervention preparation unit 112, and a database 113, and the AI ​​operator unit 111 and the intervention preparation unit 112 exist as software functions installed in ROM.

[0013] The AI ​​operator unit 111 of this embodiment has the function of periodically making calls to the terminal device 12 held by the subject 2, or receiving calls from the terminal device 12 held by the subject 2, and having a conversation with the subject 2.

[0014] The AI ​​operator unit 111 has the function of using generative artificial intelligence technology to recognize and analyze the speech of the target person 2, construct a conversation including appropriate response content to the speech of the target person 2, and synthesize the speech to be output. Here, the generative artificial intelligence refers to a model that learns on its own based on a huge amount of data and generates output data based on input of voice data, text data, or image data related to the data to be output. For example, voice data, text, or images are generated as output data. In addition, in this embodiment, the generative artificial intelligence is a large-scale language model such as a Generative Pretrained Transformer (GPT), but there are no limitations on its scale or model structure. For example, other language models combining one or more models such as a convolutional neural network (CNN), a recurrent neural network (RNN) such as a long short-term memory (LSTM) or a gated recurrent unit (GRU), or a Transformer model may be used.

[0015] The terminal 12 in this embodiment includes a landline telephone, a mobile phone (mobile phone), a tablet computer terminal with a calling function, and a personal computer with a communication function. The telecommunications network 13 in this embodiment includes a dedicated line for a landline telephone, a dedicated line for a mobile phone, and the Internet to which a personal computer is connected. These terminals 12 are connected to the server 11 via these telecommunications networks 13. For example, if the subject 2 has a landline telephone, the server 11 and the subject's landline telephone (terminal 12) are connected via a wired dedicated line (telecommunications network 13) of a telephone company. If the subject 2 has a mobile phone, the server 11 and the subject's mobile phone (terminal 12) are connected via a wireless dedicated line (telecommunications network 13) of a telephone company. Furthermore, if the subject 2 has a personal computer on which software with a calling function is installed, the server 11 and the subject's personal computer (terminal 12) are connected via the Internet (telecommunications network 13).

[0016] The AI ​​operator unit 111 of this embodiment periodically calls the terminal device 12 owned by the subject 2 or receives calls from the terminal device 12 owned by the subject 2, and various information about the subject 2 who receives the services of the dementia prevention system 1 of this embodiment is registered in advance in a database 113 provided in the server 11. The various information about the subject 2 registered in the database 113 includes the name, address, sex, age, the telephone number of the terminal device 12 owned, family composition (including spouse, children, and grandchildren), emergency contact information, career history, hobbies, medical history, regular hospital, nickname, etc.

[0017] The AI ​​operator unit 111 periodically calls the terminal 12 of the target person 2 registered in the database 113. Here, "periodically" may refer to a specific date and time, such as 9:00 AM every day, 1:00 PM on weekdays, or 5:00 PM every Sunday, or a more general date and time, such as every morning, a weekday evening, or a holiday morning. A call date and time may be registered in the database 113 after a prior agreement is reached with a specific target person 2.

[0018] In addition to making phone calls to the subject 2, the AI ​​operator unit 111 may also receive periodic calls from the subject 2 registered in the database 113. If the elderly subject 2 waits for a call from the AI ​​operator unit 111, there is a risk that the subject 2 may become a victim of special fraud such as an "it's me" scam. Therefore, the subject 2 may also be configured to periodically call the telephone number of the server 11. By having the subject 2 make regular calls, brain activity is increased, which can prevent the subject 2 from developing dementia and slow the progression of dementia. When the subject 2 periodically calls the telephone number of the server 11, if the call does not come after a specified date and time has passed, the AI ​​operator unit 111 may also be configured to call the terminal device 12 of the subject 2.

[0019] Once the AI ​​operator unit 111 connects to the terminal device 12 of the subject 2, it begins a conversation with the subject 2. There are no particular restrictions on the topic of the conversation, but various topics may be registered in the database 113 in advance, and matching topics may be extracted from the information on the subject 2 registered in the database 113. Alternatively, the content of conversations over the past few days may be reviewed, and if there is a topic that warrants follow-up on the progress since then, this may be adopted. In particular, it is preferable to manage the schedule by discussing the subject 2's upcoming schedule in addition to self-introductions by the AI ​​operator unit 111 and the subject 2. Depending on the subject 2's health condition, the topic of health may be brought up each time to check whether they are feeling unwell.

[0020] The server 11 of this embodiment includes an intervention preparation unit 112 that, when detecting an abnormality in the subject 2 from a conversation between the AI ​​operator unit 111 and the subject 2, converts the conversation between the AI ​​operator unit 111 and the subject 2 into text data and displays it on a predetermined monitor 114. The abnormality in the subject 2 is not particularly limited and may be set appropriately depending on the subject 2. For example, the abnormality may be an abnormality in the subject 2's physical condition, the subject 2 showing symptoms of dementia, the subject 2's dementia progressing, the subject 2 being a victim of a crime or at risk of becoming a victim, or the subject 2's emotional fluctuations or intonation exceeding a predetermined threshold.

[0021] To detect this type of anomaly, keywords that are strongly related to each anomaly are registered in advance in a database 113, and the intervention preparation unit 112 monitors the conversation between the AI ​​operator unit 111 and the subject 2. When a keyword appears in the conversation, the intervention preparation unit 112 determines whether or not it corresponds to an anomaly based on the combination of keywords and the frequency of their appearance.

[0022] When the intervention preparation unit 112 of this embodiment detects this type of abnormality, it converts the conversation between the AI ​​operator unit 111 and the subject 2 into text data and displays it on a predetermined monitor 114. This is so that the gerontologist-certified operator 3 can check the abnormality. The text data of the conversation displayed on the monitor 114 may be from the beginning of the conversation between the AI ​​operator unit 111 and the subject 2, or from the time when a keyword related to the abnormality appears. The gerontologist-certified operator 3 is alerted by the display of the text data of the conversation on the monitor 114, checks the content of the conversation displayed on the monitor 114, and determines whether he or she should intervene in the ongoing conversation and provide some kind of advice to the subject 2.

[0023] If the operator 3 checks the content of the conversation displayed on the monitor 114 and determines that there is no need for him / her to intervene in the ongoing conversation, the operator 3 does not take any action, and the conversation between the AI ​​operator unit 111 and the target person 2 continues as is. On the other hand, if the operator 3 checks the content of the conversation displayed on the monitor 114 and determines that he / she should intervene in the ongoing conversation and give some advice to the target person 2, the operator 3 inputs to the intervention preparation unit 112 that he / she will intervene in the ongoing conversation. The server 11 of this embodiment receives the input by the operator 3 that he / she will intervene in the conversation and stops the conversation function of the AI ​​operator unit 111. The conversation from this point on becomes a human-to-human conversation between the operator 3 and the target person 2.

[0024] System Operation Next, the information processing executed by the server will be described. Fig. 2 is a flowchart showing one embodiment of the information processing executed by the server 11 of the dementia prevention system 1 of this embodiment. In the illustrated embodiment, an example will be described in which the AI ​​operator unit 111 periodically calls the predetermined subject 2 to talk to them, but an example in which the AI ​​operator unit 111 receives a call from the predetermined subject 2 and talks to them can also be implemented in the same way.

[0025] First, in step ST1, the server 11 monitors the time and determines whether it is time (fixed time) to make a call that has been set in advance for each target person 2. Step ST1 is repeated until the fixed time arrives, and if it is determined that the fixed time has arrived, the server 11 proceeds to step ST2.

[0026] In step ST2, the AI ​​operator unit 111 calls the terminal 12 of a predetermined target person 2 registered in the database 113. If the target person 2 does not answer the call, the AI ​​operator unit 111 calls back a preset number of times. If the target person 2 still does not answer the call, the AI ​​operator unit 111 may alert the operator 3 or the like and then terminate the process.

[0027] In step ST3, the AI ​​operator unit 111 starts a conversation with the person 2 who made the call. As described above, the conversation may include self-introductions by the AI ​​operator unit 111 and the person 2, as well as topics related to the person 2's upcoming schedule and the person 2's health. The length of the conversation is not particularly limited, and the conversation may continue until the person 2 hangs up the phone. Because the conversation is initiated by the AI ​​operator unit 111, continuing the conversation as long as the person 2 has time is preferable for preventing dementia and slowing the progression of dementia.

[0028] In step ST4, when a conversation between the AI ​​operator unit 111 and the subject 2 begins, the intervention preparation unit 112 monitors the conversation and detects whether or not there is something unusual about the subject 2. Specifically, the intervention preparation unit 112 monitors the conversation between the AI ​​operator unit 111 and the subject 2 and determines whether or not there is something unusual based on whether or not keywords that have been registered in advance in the database 113 as being strongly associated with each abnormality appear in the conversation. If the intervention preparation unit 112 determines that there is no abnormality, it proceeds to step ST8, and if it determines that there is an abnormality, it proceeds to step ST5.

[0029] In step ST5, the intervention preparation unit 112 determines whether or not the intervention of the operator 3 is necessary based on the combination of keywords detected in step ST4, the frequency of appearance, etc. Note that in this step, the intervention preparation unit 112 only determines whether the intervention of the operator 3 is necessary, and the operator 3 himself makes the final decision (step ST7). If the intervention preparation unit 112 determines that the intervention of the operator 3 is not necessary, the process proceeds to step ST8, and if it determines that the intervention of the operator 3 is necessary, the process proceeds to step ST6.

[0030] In step ST6, the intervention preparation unit 112 converts the conversation between the AI ​​operator unit 111 and the subject 2 into text data and displays it on a predetermined monitor 114. At this time, a separate warning may be given to the operator 3. The operator 3, who is a gerontologist, checks the content of the conversation displayed on the monitor 114 and determines whether or not he or she should intervene in the ongoing conversation and give some advice to the subject 2. If the operator 3 checks the content of the conversation displayed on the monitor 114 and determines that there is no need for him or her to intervene in the ongoing conversation, the operator 3 does nothing and the process proceeds to step ST8, where the conversation between the AI ​​operator unit 111 and the subject 2 continues.

[0031] On the other hand, if the operator 3 decides that he or she should intervene in the ongoing conversation and give some advice to the target person 2, the operator 3 inputs to the intervention preparation unit 112 that he or she will intervene in the ongoing conversation. This causes the process to proceed to step ST9. In step ST9, the server 11 stops the conversation function of the AI ​​operator unit 111. The conversation from this point on becomes a human-to-human conversation between the operator 3 and the target person 2.

[0032] If no keyword related to an abnormality is detected in step ST4, if it is determined in step ST5 that there is no need for intervention by the operator 3, or if the operator 3 determines in step ST7 that there is no need for intervention, the process proceeds to step ST8, where it is determined whether or not the conversation has ended. Whether or not the conversation has ended may be determined based on whether or not a preset conversation time has elapsed, in addition to whether or not the subject 2 has made an utterance to the effect that the conversation is to end.

[0033] If the conversation is finished in step ST8, the process proceeds to step ST9, where the server 11 stops the conversation function by the AI ​​operator unit 111. If the conversation is not finished in step ST8, the process returns to step ST4, and the processes from step ST4 onwards are executed again.

[0034] 《Examples of conversations with the target person: everyday conversations》 Next, an example of an actual conversation between the AI ​​operator unit 111 and the subject 2 will be described. FIG. 3 is a time chart showing an example of a conversation carried out in the dementia prevention system 1 of this embodiment. An utterance from the subject 2 is indicated by X, and an utterance from the AI ​​operator unit 111 is indicated by Y. The conversation example shown in FIG. 3 is an ordinary conversation that takes place in an everyday morning, in which the AI ​​operator unit 111 calls the subject 2 at a fixed time in the morning.

[0035] This is an example of a 10-minute light conversation to start the day, including checking whether the person has had breakfast and what their plans are for the day. By providing casual morning conversations like this to elderly people living alone, not only can they activate their brains, but they can also prevent problems in their daily lives by accurately communicating their plans for the day.

[0036] In the utterance Y1 from the AI ​​operator unit 111, by calling the target person 2 by their first name instead of their last name, or by personifying the AI ​​operator unit 111 itself to convey their name, it is possible to reduce the psychological distance between the target person 2 and the AI ​​operator unit 111. In addition, by constructing a conversational sentence that includes information on the current local weather, it becomes a suitable topic to start the day and also helps to spark a conversation with the target person 2.

[0037] In utterance Y2 from the AI ​​operator unit 111, the schedule of the target person 2 is managed based on the previous conversations recorded in the database 113. For example, the AI ​​operator unit 111 checks with the other person whether there have been any changes to the schedule, confirms the schedule, informs the other person of upcoming plans, and encourages them to start preparing. This allows the target person 2 to remember today's plans that they had forgotten, as in utterance X2 from the target person 2, and also allows them to start preparations with plenty of time to spare.

[0038] In utterance Y3 from the AI ​​operator unit 111, by repeating and confirming utterance X2 from target person 2, it is possible to further impress upon target person 2 what today's schedule is. Furthermore, the AI ​​operator unit 111 praises target person 2, checks whether he / she knows the location of the hospital, explains the benefits of walking, and tells him / her how far he / she has to walk to the hospital and how much further he / she still has to walk. Furthermore, in utterance Y4 from the AI ​​operator unit 111, it confirms target person 2's new schedule.

[0039] In speech Y5 from the AI ​​operator unit 111, hobbies and unfinished business are extracted from past conversations recorded in the database 113 and brought up as topics. If there are no plans after the hospital visit, suggestions are made on how to utilize the time. In speech Y6 from the AI ​​operator unit 111, the call is ended in a way that makes the patient look forward to the next call.

[0040] Example of a conversation with a target: Conversation regarding anomaly detection (victim of special fraud) Fig. 4 is a time chart showing another example of a conversation carried out in the dementia prevention system 1 of this embodiment. The conversation example shown in Fig. 4 is an example in which, in an ordinary conversation, it is detected that a person has been or is at risk of being victimized by a special fraud (or is currently being victimized), so in order to prevent the person from becoming a victim of a special fraud, the system reacts to withdrawals from a bank or post office, checks the situation in detail, and if there is any danger, reports to the police.

[0041] When the AI ​​operator unit 111 inquires about the actions of subject 2 yesterday in utterance Y11, it is confirmed in utterance X11 that subject 2 went alone to a nearby post office. Here, "post office" included in utterance X11 from subject 2 is a keyword that has been recorded in advance in database 113 as being strongly related to abnormalities, and therefore it is determined in step ST4 shown in Figure 2 that there is an abnormality (keyword). Therefore, in the subsequent utterances Y12 to Y14 from the AI ​​operator unit 111, topics related to the post office are offered in order to hear more detailed information about the "post office."

[0042] Based on the utterances X12-14 from the subject 2 in response to the utterances Y12-Y14 from the AI ​​operator unit 111, the intervention preparation unit 112 determines whether or not the intervention of the operator 3 is necessary in step ST5 shown in Figure 2. If it determines that the intervention of the operator 3 is necessary, the previous utterances Y11-Y14 and the utterances X11-X14 in response to these utterances are converted into text data and displayed on the monitor 114.

[0043] Operator 3, having checked the conversation displayed on monitor 114, is able to confirm that subject 2 has fallen victim to a specialized fraud, has withdrawn money at the post office yesterday, and plans to hand it over to the fraud suspect around noon today, so he intervenes in the conversation that has been taking place up to that point through AI operator unit 111, and in utterance Y16, operator 3, a human with gerontology qualifications, converses with subject 2. Here, subject 2 explains the current situation in order to prevent the subject from becoming a victim of specialized fraud.

[0044] Example of conversation with the subject: Conversation regarding abnormality detection (illness) Fig. 5 is a time chart showing yet another example of a conversation carried out in the dementia prevention system 1 of this embodiment. The conversation example shown in Fig. 5 is an example in which, during an ordinary conversation, it is detected that the subject 2 has not been feeling well for the past few days, and the AI ​​operator unit 111 recommends that the subject 2 notify a family member or go to the hospital. This conversation is intended to build a relationship in which the AI ​​operator unit 111 is there for the subject 2 and offers advice when the subject 2 is feeling vaguely depressed or lethargic.

[0045] When the AI ​​operator unit 111 asks about the subject 2's physical condition today in utterance Y21, it is confirmed that the subject 2 feels feverish and fatigued in utterance X21. Here, the words "feverish" and "fatigued" contained in utterance X21 from the subject 2 are keywords that have been recorded in advance in the database 113 as being strongly related to abnormalities, and therefore it is determined in step ST4 shown in Figure 2 that there is an abnormality (keyword). Therefore, in the subsequent utterances Y22 to Y27 from the AI ​​operator unit 111, topics related to the subject 2's physical condition are provided in order to obtain more detailed information about the extent of the subject 2's poor physical condition.

[0046] Based on the utterances X22-26 from the subject 2 in response to the utterances Y22-Y26 from the AI ​​operator unit 111, the intervention preparation unit 112 determines whether or not the intervention of the operator 3 is necessary in step ST5 shown in Fig. 2. In the example shown in Fig. 5, it is determined that the intervention of the operator 3 is not necessary, and instead the AI ​​operator unit 111 makes various suggestions and confirmations to the subject 2.

[0047] For example, in utterance Y22 from the AI ​​operator unit 111, it checks whether the patient is feeling unwell enough to inform a family member such as a son, and in utterance Y23, it checks the date on which the patient's condition worsened. In utterance Y24 from the AI ​​operator unit 111, it recommends going to a regular hospital, and if it is confirmed from utterance X24 of subject 2 that the patient is willing to go to the hospital, it calls a taxi for the patient in utterances Y25 and Y26.

[0048] <System Effects> As described above, the dementia prevention system 1 of this embodiment includes a server 11 equipped with an AI operator unit 111 that conducts conversations using artificial intelligence, a terminal 12 owned by a subject 2 whose dementia should be prevented or whose dementia progression should be slowed, and a telecommunications network 13 that connects the server 11 and the terminal 12. The AI ​​operator unit 111 periodically calls the subject 2's terminal 12 or receives calls from the subject's terminal, and converses with the subject 2, thereby stimulating brain activity in the subject 2, relieving stress through conversation, and building a trusting relationship with the caller to stabilize the subject's mind. As a result, the subject 2 can be prevented from developing dementia and the progression of dementia can be slowed.

[0049] Furthermore, in the dementia prevention system 1 of this embodiment, the server 11 further includes an intervention preparation unit 112 that, when detecting from the conversation between the AI ​​operator unit 111 and the subject 2 that any of the following is detected: that there is an abnormality in the physical condition of the subject 2; that the subject 2 has symptoms of dementia; that the subject 2's dementia is progressing; that the subject 2 has been or is at risk of being a victim of a crime; or that the subject 2's emotional ups and downs exceed a predetermined threshold, converts the conversation between the AI ​​operator unit 111 and the subject 2 into text data and displays it on a predetermined monitor 114. Therefore, by detecting signs of an abnormality in the subject 2 and displaying them on the monitor, it is possible to alert the human operator 3. As a result, the abnormality can be dealt with by the final decision of the human operator 3, not by the AI ​​operator unit 111.

[0050] In addition, in the dementia prevention system 1 of this embodiment, when an operator 3 with gerontology qualifications inputs to the intervention preparation unit 112 that he or she will intervene in the conversation between the AI ​​operator unit 111 and the subject 2, the server 11 stops the conversation function of the AI ​​operator unit 111, so that the conversation can transition to the conversation between the operator 3 and the subject 2. [Explanation of symbols]

[0051] 1. Dementia prevention system 11...Server 111…AI Operator Section 112…Intervention Preparation Department 113...Database 114...Monitor 12...Terminal 13...Telecommunications network 2. Target 3...Operator

Claims

1. a server having an AI operator unit that conducts conversations using artificial intelligence; A terminal device held by a subject for whom dementia should be prevented or dementia progression should be suppressed; a telecommunications network connecting the server and the terminal; The AI ​​operator unit is a dementia prevention system that periodically calls the subject's terminal device or receives calls from the subject's terminal device and converses with the subject.

2. The server 2. The dementia prevention system of claim 1, further comprising an intervention preparation unit that, when detecting from the conversation between the AI ​​operator unit and the subject that there is an abnormality in the subject's physical condition, that the subject has symptoms of dementia, that the subject's dementia is progressing, that the subject has been the victim of a crime or is at risk of being the victim, or that the subject's emotional ups and downs exceed a predetermined threshold, converts the conversation between the AI ​​operator unit and the subject into text data and displays it on a predetermined monitor.

3. The server The dementia prevention system of claim 2, wherein the conversation function of the AI ​​operator unit is stopped when an operator with gerontology qualifications inputs to the intervention preparation unit that he or she will intervene in the conversation between the AI ​​operator unit and the subject.

4. The AI ​​operator unit selects at least one of the following topics for the conversation with the subject: a topic of self-introduction by the AI ​​operator unit; and a topic of schedule management for the subject. The dementia prevention system according to any one of claims 1 to 3.

Images