Harassment information providing device, harassment information providing method and computer program

Abstract

To provide technology that can present convincing information that leads to suppression of harassment to a person who has made a statement that raises concerns about harassment in a conversation.SOLUTION: A harassment information providing device is equipped with a transcription unit, a determination unit, and an output unit. The transcription unit converts speaker's remarks into text from voice data of conversation. The determination unit calculates a degree of negativity and positivity of the conversation by analyzing the text data representing the textualized statements, and determines a degree of harassment of the conversation using the negativity and positivity. The output unit outputs harassment information including information representing the degree of harassment.SELECTED DRAWING: Figure 11

Description

[Technical field]

[0001] The present disclosure provides a technique for suppressing harassment. [Background technology]

[0002] As shown by the enforcement of the so-called Power Harassment Prevention Act, there is a growing awareness of the need to suppress harassment in the workplace etc. Patent Document 1 (JP 2020-9238 A) and Patent Document 2 (JP 2023-9563 A) disclose a technology for detecting and notifying harassment using NG words for harassment detection in order to prevent harassment. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2020-9238 A [Patent Document 2] JP 2023-9563 A Summary of the Invention [Problem to be solved by the invention]

[0004] Even if a similar statement that raises concerns about harassment is made in a conversation, the recipient of the statement may or may not feel that they have been harassed. For this reason, it is difficult to accurately detect harassment in a conversation based solely on statements that raise concerns about harassment. Furthermore, since people who make statements that raise concerns about harassment (hereinafter referred to as harassing statements) often do not realize that they have made harassing statements, they may not be satisfied if they are simply notified that harassment has been detected through a harassment presence / absence determination. For this reason, it is thought that simply notifying that harassment has been detected through such a harassment presence / absence determination may not lead to the prevention of harassment.

[0005] The present disclosure has been devised to solve the above problems. That is, the main objective of the present disclosure is to provide a technology that can present convincing information that leads to the suppression of harassment to a person who has made a statement in a conversation that may be considered harassment. [Means for solving the problem]

[0006] In order to achieve the above object, the harassment information providing device according to the present disclosure includes, as one aspect thereof, a transcription unit that converts the speaker's utterances into text from the audio data in the conversation; a determination unit that calculates a negative degree and a positive degree of a conversation by analyzing text data representing a converted remark, and determines a harassment degree of the conversation using the negative degree and the positive degree; an output unit for outputting harassment information including information indicating the determined degree of harassment; Equipped with.

[0007] In addition, one aspect of the method for providing information about harassment in the present disclosure is as follows: By computer, Convert the speaker's remarks into text from the audio data of the conversation, Calculating the degree of negativity and positivity of the conversation by analyzing the text data representing the transcribed remarks, and determining the degree of harassment of the conversation using the degree of negativity and positivity; Harassment information including information indicating the degree of harassment is output.

[0008] Furthermore, in one aspect, the computer program of the present disclosure includes: A process of converting the speaker's utterances into text from the voice data in the conversation; A process of calculating the degree of negativity and positivity of a conversation by analyzing text data representing the converted remarks and determining the degree of harassment of the conversation using the degree of negativity and positivity; A process for outputting harassment information including information indicating the degree of harassment; to be executed by the computer. Effect of the Invention

[0009] According to the present disclosure, convincing information that will lead to the prevention of harassment can be presented to a person who makes a statement in a conversation that raises concerns of harassment. [Brief description of the drawings]

[0010] [Figure 1] 1 is a diagram illustrating a configuration example of a harassment information providing device (information providing device) in the present disclosure. [Diagram 2] FIG. 2 is a diagram illustrating an example of text data obtained by converting voice data into text through a transcription process. [Diagram 3] FIG. 13 is a diagram showing an example of a negative word. [Figure 4] FIG. 13 is a diagram showing an example of a positive word. [Diagram 5] FIG. 13 is a diagram showing an example of a positive word. [Figure 6] FIG. 13 is a diagram illustrating an example of a display of harassment information on a display device. [Figure 7] FIG. 11 is a diagram illustrating another example of a display of harassment information. [Figure 8] FIG. 13 is a diagram showing yet another example of a display of harassment information. [Figure 9] 10 is a flowchart illustrating an example of the operation of the harassment information providing device (information providing device). [Figure 10] FIG. 13 is a diagram illustrating yet another example of displaying harassment information. [Figure 11] FIG. 13 is a diagram illustrating another example of the configuration of the harassment information providing device (information providing device). [Figure 12] 13 is a flowchart illustrating another example of the operation of the harassment information providing device (information providing device). DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] Hereinafter, embodiments of the present disclosure will be described with reference to the drawings.

[0012] First Embodiment The harassment information providing system of the first embodiment of the present disclosure is a system for suppressing harassment in the workplace and the like, focusing on conversation. This harassment information providing system has a function for determining the degree of harassment of remarks in conversation, and a function for outputting harassment information including the degree of harassment. Harassment is what is called harassment, and is an act of causing physical and mental pain, such as making the other person feel uncomfortable or disadvantaged, and there are multiple types. Some types of harassment include power harassment (power harassment), sexual harassment (sexual harassment), maternity harassment (matahara), and moral harassment (morahara). Here, the type of harassment to be suppressed is not limited.

[0013] The harassment information output by the harassment information providing system of the first embodiment includes information indicating the degree of harassment in a conversation. This harassment information is provided (output) at least to a person who is determined to have made a harassing remark (hereinafter, also referred to as a harasser).

[0014] The degree of harassment contained in the harassment information is determined (calculated) taking into consideration the positive degree as well as the negative degree associated with the harassing remarks in the conversation. That is, the inventor has noted that the degree of harassment does not necessarily increase as the number of negative remarks (remarks that lead to harassment (harassing remarks)) in the conversation increases. The inventor has considered that the reason for this is that a conversation may contain positive remarks (remarks that reduce harassment), and the inclusion of positive remarks in a conversation reduces the harassment degree of the conversation. From this, the inventor has considered that the accuracy of the degree of harassment can be increased by using the positive degree calculated from the number of positive remarks in addition to the negative degree, rather than determining the degree of harassment only from the negative degree calculated from the number of negative remarks.

[0015] In the harassment information providing system of the first embodiment, the harassment speaker is not simply notified that harassment has been detected by the presence / absence of harassment judgment, but instead, as described above, the harassment information including information indicating the degree of harassment in the conversation is output. It is considered that the harassment speaker is often unaware that he / she has made a harassing remark. In addition, there is a tendency for harassment to gradually escalate from a low level to a high level. For this reason, the harassment information providing system of the first embodiment notifies (outputs) the harassment speaker of harassment information from a low level to a high level, thereby making the harassment speaker aware of his / her own harassment remark before the level of harassment becomes high. In addition, by notifying the harassment speaker of the information indicating the level of harassment, the harassment speaker is made aware of the harassment remark with a sense of conviction. In other words, there is a problem that even if the harassment speaker is warned about harassment when the level of harassment of the harassment remark is low, the harassment speaker does not feel convinced, which makes it difficult to suppress the harassment. In contrast, the harassment information providing system of the first embodiment is expected to be able to convey to the harasser that "the comment is harassing, although not severe," by notifying the harasser of the degree of harassment.

[0016] The configuration of the harassment information providing system of the first embodiment will be described below. The harassment information providing system 1 of the first embodiment includes a harassment information providing device (hereinafter, also abbreviated as information providing device) 2, as shown in Fig. 1. The information providing device 2 is connected to a microphone 6.

[0017] The microphone 6 has a function of collecting the voice of a conversation and converting the voice into voice data, which is an electrical signal, and outputting it. As described above, the harassment information provision system 1 of the first embodiment focuses on harassment in conversation. A conversation may be conducted directly face-to-face, or may be conducted using an information device (including a telephone) rather than face-to-face, such as a telephone call or an online conference. Either case can be handled here. For example, when collecting the voice of a conversation conducted face-to-face, the microphone 6 is placed in a place where the utterance of the conversation in the workplace or the like that is the target of harassment suppression can be collected. In this case, the microphone 6 may be a standalone device or may be built into an information device such as a smartphone. Also, when collecting the voice of a conversation conducted using an information device rather than face-to-face, such as a telephone call or an online conference, a microphone provided in the information device functions as the microphone 6 used in this system.

[0018] The information providing device 2 is a computer device, and includes an arithmetic unit 20 and a storage device 30. The storage device 30 includes a storage medium for storing data and a computer program (hereinafter also referred to as a program) 31. There are multiple types of storage devices, such as a magnetic disk device and a semiconductor memory element, and there are multiple types of semiconductor memory elements, such as a RAM (Random Access Memory) and a ROM (Read Only Memory). The information providing device 2 includes multiple storage devices with different uses, but the types and numbers of the storage devices are not limited here, and the description thereof will be omitted. Also, the multiple storage devices included in the information providing device 2 will be collectively referred to as the storage device 30 without distinction.

[0019] The arithmetic device 20 is configured with a processor such as a CPU (Central Processing Unit) or a GPU (Graphics Processing Unit). The arithmetic device 20 can have various functions based on a program 31 by reading and executing the program 31 stored in a storage device 30. Here, the arithmetic device 20 includes an acquisition unit 21, a transcription unit 22, a determination unit 23, and an output unit 24 as functional units related to the provision of harassment information.

[0020] The acquisition unit 21 acquires voice data of the conversation collected by the microphone 6. The following modes are conceivable as modes in which the voice data of the conversation by the microphone 6 is provided to the information providing device 2. In one of the modes, for example, the microphone 6 or an information device equipped with the microphone 6 is connected to the information providing device 2, and the voice data is provided (transmitted) from the microphone 6 or the information device to the information providing device 2 in real time. In another mode, the microphone 6 is built into an information device such as a smartphone, and the voice data of the conversation by the microphone 6 is recorded by a recording function of the information device, and the recorded voice data is provided from the information device to the information providing device 2. In the first embodiment, the mode in which the voice data by the microphone 6 is provided to the information providing device 2 is not limited.

[0021] The acquisition unit 21 acquires voice data of a conversation through the microphone 6 provided to the information providing device 2. The acquisition unit 21 then stores the acquired voice data in the storage device 30. In addition, when the information providing device 2 is connected to a database 4 that is a storage device, the acquisition unit 21 may store the acquired voice data in the database 4. The voice data stored in the storage device 30 or the database 40 is, for example, one file (storage unit) of voice data from the beginning to the end of a conversation. Information on the beginning and end of the conversation is input to the information providing device 2 by, for example, an operator of the information providing device 2. In addition, the file of the voice data stored in the storage device 30 or the database 4 is associated with file identification information that identifies the file of the voice data, time information that indicates the date and time when the conversation took place, and the like. In addition, when the voice data is sequentially provided from the microphone 6 in real time, for example, the voice data is temporarily stored in sequence, and then compiled into one file of voice data as described above and stored in the storage device 30.

[0022] The transcription unit 22 executes a transcription process for converting the speech of the speaker of the conversation acquired by the acquisition unit 21 into text. Here, the conversion into text means converting the speech data into text data (character data). There are various methods for the transcription process for converting speech data into text data, and the transcription process method is not limited here, and the description thereof is omitted. However, the transcription process here is not limited to simply converting speech data into text data, but also includes a process for distinguishing speech of multiple speakers who are conversing from one speaker to another. An example of the contents of a conversation that has been converted into text by such a transcription process is shown in FIG. 2. In the example of FIG. 2, the contents of the speech are expressed in characters in chronological order from the top, and the identification information of the speaker representing the speaker who made the speech is associated with the character string of the speech.

[0023] The transcription process by the transcription unit 22 may be performed after the conversation has ended, or in real time while the conversation is taking place. When the transcription process is performed after the conversation has ended, the file of the audio data to be transcribed is stored in the storage device 30 or the database 4. For this reason, the transcription unit 22 reads out the file of the audio data to be transcribed from the storage device 30 or the database 4, and converts the read audio data into text data. Examples of the timing for executing such a transcription process include when a command is received from an operator, or when it is detected that the audio data file has been stored in the storage device 30 or the database 4.

[0024] There is a case where the voice data of the conversation is transmitted to the information providing device 2 in real time, and setting information for executing the transcription process in real time is given to the information providing device 2. In this case, the transcription unit 22 converts the transmitted voice data into text data in sequence. In this case, the text data converted by the transcription process is associated with the original voice data that was subjected to the transcription process and was transmitted in real time and temporarily stored.

[0025] When the transcription unit 22 completes the transcription process of the conversation, it stores the text data file obtained by the transcription process in the storage device 30 or database 4 in a manner associated with the original audio data file that underwent the process.

[0026] The judgment unit 23 judges the harassment level of a conversation for each speaker using text data representing the contents of the conversation. When judging the harassment level of a conversation, the judgment unit 23 uses not only the negative degree of the conversation but also the positive degree. The reason for this is that, as described above, a conversation may contain positive remarks (remarks that reduce harassment), and the inventor has noticed that when a positive remark is included in a conversation, the harassment level of the conversation decreases. From this, the inventor has considered that the accuracy of the harassment level can be increased by using the positive degree (positive remarks) in addition to the negative degree, rather than judging the harassment level only based on the negative degree (negative remarks).

[0027] For this reason, the judgment unit 23 first calculates the degree of negativity and the degree of positivity by analyzing the text data representing the contents of the conversation.The judgment unit 23 then uses the detected degrees of negativity and positivity to judge the degree of harassment of the conversation.

[0028] The following method is an example of a method for calculating the degree of negativity and the degree of positivity in a conversation. For example, data on negative words and positive words as shown in Figs. 3 to 5 is stored in advance in the storage device 30 of the information providing device 2. By referring to such data, negative words (negative statements) and positive words (positive statements) included in the conversation are extracted from the text data of the conversation, and the negative words and positive words are counted by the determination unit 23. Then, the negative degree is calculated by multiplying the count value of the negative words by a coefficient for negativity. Also, the positive degree is calculated by multiplying the count value of the positive words by a coefficient for positivity. The coefficients for negativity and the coefficient for positivity change according to the length of the conversation in which the negative words and the positive words are counted. In other words, since the length of the conversation in which the negative words and the positive words are counted varies, the length of the conversation is normalized by the coefficients for negativity and the coefficient for positivity. In addition, the magnitude relationship between the negative and positive coefficients is set taking into consideration the difference in the magnitude of the impact that negative comments have on people and the impact that positive comments have on people. The difference in the magnitude of the impact that negative comments have on people and the impact that positive comments have on people can be obtained, for example, by experiments or simulations.

[0029] When the determination unit 23 calculates the degree of negativity and the degree of positivity after the conversation ends, the degree of negativity and the degree of positivity for one conversation (one file of audio data) from the beginning to the end of the conversation are calculated. When the acquisition and transcription of audio data is performed in real time during a conversation, for example, the negative words and positive words are counted in real time by the determination unit 23. Then, the degree of negativity and the degree of positivity for the conversation from the beginning of the conversation to the time when the process of counting the negative words and positive words is performed is calculated.

[0030] Incidentally, it is considered that the way of feeling when receiving words, especially when receiving positive words, differs depending on the situation and the person. In consideration of this, the following method may be used as a method of calculating the negative degree and the positive degree, taking into account the statements before and after the negative and positive words. For example, 10 points are assigned as a standard value for each positive word. Here, if it is determined that the positive impact of a certain positive word is lower than the standard based on the statements before and after, the points assigned to the positive word are lowered by the amount of the low impact, for example, 7 points. In this way, the positive impact is quantified for each positive word. Furthermore, by adding up the points of each positive word extracted from the conversation, the number of positive words in the conversation and the points that quantify their impact are calculated. Points are calculated similarly for negative words. The negative degree and the positive degree may be calculated using the points for the positive words calculated in this way, the points for the negative words, and the negative and positive coefficients described above.

[0031] The judgment unit 23 further calculates the degree of harassment in the conversation using the negative degree and positive degree calculated as described above. One example of a method for calculating the degree of harassment is a method for calculating the degree of harassment by subtracting the value of the positive degree from the value of the negative degree. The degree of harassment may be expressed by a numerical value calculated by such a calculation method, or may be expressed by a rank (level) that divides the degree of the degree into a plurality of categories. The manner of expressing the degree is appropriately set by the system designer, taking into consideration, for example, the ease of conveying the degree of harassment.

[0032] The determination unit 23 determines (calculates) the degree of harassment for each speaker of the conversation using the negative and positive degrees calculated as described above. The determination unit 23 also stores the information on the degree of harassment for each speaker of the conversation thus determined in the storage device 30 or the database 4 in association with the corresponding voice data file.

[0033] The output unit 24 outputs the harassment information. The harassment information is information including information representing the degree of harassment in a conversation. This harassment information is output at least to the person who made the harassing remark (the harasser). In other words, the harassment information is information having one of the purposes of making the harasser realize that he or she has made a harassing remark. Whether or not a speaker in a conversation is a harasser is determined, for example, by using the harassment degree calculated for each speaker. For example, a threshold value for determining whether or not a speaker is a harasser using the harassment degree is set in advance, and a speaker whose harassment degree is equal to or greater than the threshold is determined to be a harasser.

[0034] The output unit 24 outputs, to at least the harasser who has been determined as such, harassment information including information indicating the degree of harassment in the conversation of the harasser. For example, email is used to output the harassment information. When email is used, email address information linking the speaker of the conversation (speaker identification information) with the email address information of the speaker is stored in advance in the storage device 30. The output unit 24 extracts the email address of the harasser from the email address information using the speaker identification information corresponding to the harasser, and outputs the harassment information to the harasser by email using the address. Note that the email may contain the harassment information itself or may be attached thereto, or may contain link information indicating the storage location of, for example, database 4 where the harassment information is stored.

[0035] In addition, when the degree of harassment in a conversation is judged after the conversation ends, the timing at which the output unit 24 outputs the harassment information may be, for example, a timing set by a system setter after the degree of harassment is judged. In addition, when the degree of harassment is judged (calculated) in real time during a conversation, the harassment information can be output during the conversation. However, if the conversation time from the start of the conversation is short, there is a concern that unreliable information on the degree of harassment may be output. In consideration of this, for example, an output condition for the harassment information during a conversation is set in advance, and the harassment information is output during the conversation when the output condition is satisfied. This is another example of the timing at which the output unit 24 outputs the harassment information.

[0036] The harassment information output by the output unit 24 is displayed, for example, on a display device (not shown) of the terminal device 7 operated by the harasser. FIG. 6 shows an example of the display of the harassment information. In the example of FIG. 6, the degree of harassment of the harasser in the conversation that caused the harassment information to be output is displayed, as well as the negative and positive degrees. In addition, in the example of FIG. 6, text data of the conversation that caused the harassment information to be output is also displayed. Furthermore, in the example of FIG. 6, displays (marks in the example of FIG. 6) are displayed to clarify statements that include negative words and statements that include positive words. Furthermore, as shown in FIG. 7, explanations of terms such as positive statements and negative statements may be popped up on the display screen that displays the harassment information. When the displays shown in FIG. 6 and FIG. 7 are made on the display device by the display control operation of the terminal device 7, the harassment information includes the following information in addition to the information on the degree of harassment. In other words, the harassment information includes information on the degree of harassment, the negative degree and the positive degree, text data for each speaker of the conversation, and information for identifying negative and positive words included in the text data. Furthermore, the harassment information may include address information (link information) for connecting to, for example, a web page showing an explanation of the term.

[0037] Furthermore, information showing the analysis results of the conversation may be associated with the harassment information. In this case, for example, display control may be performed to transition from a display screen of the harassment information as shown in Fig. 6 to a screen showing the analysis results of the conversation as shown in Fig. 8. In the example of Fig. 8, the statements made by the harasser in the conversation are classified into positive statements, negative statements, and other statements (others), and the proportion of these statements in the conversation is displayed. Furthermore, the information on the analysis results of the conversation may include information on comments showing things to be careful about in the conversation, and in this case, such information (comments) may also be displayed.

[0038] The harassment information providing system 1 of the first embodiment is configured as described above. An example of the operation of the information providing device 2 of the first embodiment will be described below with reference to Fig. 9. Fig. 9 is a flowchart illustrating an example of the operation of the information providing device 2.

[0039] For example, with respect to a conversation that is the subject of a harassment check, first, the acquisition unit 21 acquires voice data of the conversation collected by the microphone 6 (step 101). As a result, the transcription unit 22 executes a transcription process that converts the utterances of the conversation from the acquired voice data into text (step 102).

[0040] Then, when the conversational remarks are converted into text data by the transcription process, the determination unit 23 analyzes the text data and calculates the degree of positivity and the degree of negativity. Furthermore, the determination unit 23 calculates the degree of harassment using the calculated degrees of negativity and positivity (step 103).

[0041] Thereafter, the output unit 24 outputs the harassment information including the degree of harassment (step 104). For example, this harassment information is output at least to the harasser, thereby making the harasser aware that he or she has made a statement that may be considered harassment.

[0042] The information providing device 2 of the first embodiment constituting the harassment information providing system 1 has the configuration described above. That is, the information providing device 2 has a configuration for determining the degree of harassment in a conversation by considering not only negative remarks that may be considered harassment but also positive remarks. This allows the information providing device 2 to increase the accuracy of the determination result of the harassment degree compared to a case where the harassment degree in a conversation is determined (calculated) using only negative remarks, and to increase the reliability of the determination (calculation) of the harassment degree. In other words, the information providing device 2 can present convincing information that leads to the suppression of harassment to a person who has made a remark that may be considered harassment in a conversation (harasser).

[0043] Furthermore, the information providing device 2 does not simply output the detection of harassment based on the presence or absence of harassment, but outputs harassment information including information indicating the degree of harassment. Therefore, even when the degree of harassment is light, the information providing device 2 is expected to be able to make the harasser aware of the harassing remark without causing discomfort by outputting harassment information including information on the degree of harassment. Furthermore, if the harasser can be made aware of the harassing remark while the degree of harassment is still light, it is believed that the harassment can be suppressed before the degree of harassment by the harasser worsens. In other words, the information providing device 2 can suppress the spread of the damage caused by harassment.

[0044] In the above example, the terminal device 7 is given as an example of the output destination of the output unit 24, but for example, the output destination may also be a printer 8 as represented by the dotted line in Figure 1, and a paper medium 9 on which the harassment information is printed by the printer 8 may be provided to the person making the harassment.

[0045] In the above example, the harassment information is output to the harasser, but the harassment information may be output to all speakers in the conversation. In the case of a workplace, the harassment information may also be output to the harasser's superior. In this way, the output destination of the harassment information may be set appropriately, taking into consideration the suppression of harassment.

[0046] <Second embodiment> A second embodiment of the present disclosure will be described below. In the description of the second embodiment, the same reference numerals are used to designate components that are the same as those in the harassment information providing system of the first embodiment, and redundant description of the common components will be omitted.

[0047] In the second embodiment, the judgment unit 23 in the information providing device (harassment information providing device) 2 calculates the degree of negativity and the degree of positivity, respectively, using the following information in addition to the information on the number of negative words and positive words in a conversation as described in the first embodiment.

[0048] That is, even when the same negative words are used, the degree of negativity felt by the receiver of the utterance differs depending on the tone of the utterance, such as the volume and roughness of the voice, and the relationship between the speakers of the conversation (for example, the relationship between a boss and a subordinate, or between colleagues, or between a store clerk and a customer). Similarly, even when the same positive words are used, the degree of positivity felt by the receiver of the utterance differs depending on the tone of the utterance and the relationship between the speakers of the conversation. Taking this into consideration, in the second embodiment, the degree of negativity calculated from the number of negative words and the degree of positivity calculated from the number of positive words are corrected by the judgment unit 23 using the following information. The judgment unit 23 determines the corrected degrees of negativity and positivity as the degrees of negativity and positivity, respectively.

[0049] For example, as described above, the way in which the receiver (the person receiving the speech) feels the speech from the conversation partner varies depending on the tone of speech and the relationship between the speakers of the conversation. For this reason, it is conceivable to correct the degree of negativity and the degree of positivity using information on the tone of speech and information on the relationship between the speakers of the conversation that can be acquired from the voice data. In this case, for example, relationship data between elements whose values ​​change in speech, such as the volume of voice and the speed of speech, and the correction amount for correcting the degree of negativity and the degree of positivity according to the amount of change in the value of the element is obtained in advance by experiment or simulation, and stored in the storage device 30, etc. In addition, relationship data between the relationship between the speakers of the conversation and the correction amount of the degree of negativity and the degree of positivity is also obtained in advance by experiment or simulation, and stored in the storage device 30, etc. Using such relationship data, information on the values ​​of elements such as the volume of the speaker's voice and the speed of speech acquired (detected) from the voice data of the conversation, and information representing the relationship between the speakers, the degree of negativity and the degree of positivity in the conversation for each speaker are corrected by the judgment unit 23. Note that the information representing the relationship between the speakers is input to the information providing device 2 by, for example, the operator of the information providing device 2. Furthermore, the relationship between the speakers is not limited to the above-mentioned examples, and examples include relationships based on the length of their relationship, relationships based on age difference, the relationship between a full-time employee and a contract employee, and relationships based on being of the same sex.

[0050] In addition, it is considered that the difference in the way the receiver feels due to the speech of the conversation partner is reflected in the magnitude of the change in the speaker's emotion when the speech is received. From this, for example, it is considered that the degree of negativity and the degree of positivity are corrected using information on the magnitude of the change in the emotions of the speakers during the conversation. For example, information on the magnitude of the change in the emotions of the speakers during the conversation is acquired, and the degree of negativity and the degree of positivity are corrected by the judgment unit 23 according to the difference between the acquired magnitude of the change in emotions and the magnitude of the change in a predetermined standard. As an example of a method of acquiring information on the magnitude of the change in the emotions of the speakers, a method of acquiring it from the change in facial expression in a face image obtained by photographing the faces of the speakers can be given. In addition, as another example of a method of acquiring information on the magnitude of the change in the emotions of the speakers, a method of acquiring it from the change in heart rate obtained from the change in skin color acquired from the photographed image of the faces of the speakers can be given. Furthermore, there is also a method of acquiring information on the magnitude of the change in emotions from the change in heart rate acquired by the wearable device worn by the speakers. In this way, there are multiple types of methods for acquiring information on the magnitude of the change in the emotions of the speakers. Here, information on the magnitude of change in emotion of the speaker acquired by any one of such methods may be used, or multiple pieces of information on the magnitude of change in emotion acquired by multiple acquisition methods may be used. When multiple pieces of information on the magnitude of change in emotion are used, for example, a statistical process is performed in which an average value of the multiple pieces of information is calculated and the calculated value is used.

[0051] Furthermore, it is considered that the speaker's interlocutor's appearance, such as clothing and hairstyle, and the interlocutor's gestures also affect how the receiver feels about the speaker's remarks. For this reason, it is considered that the degree of negativity and positivity can be corrected using information on the interlocutor's appearance, such as clothing and hairstyle, and the interlocutor's gestures. In this case, for example, information on the speakers' appearance and gestures is acquired from a photographed image of the speakers, and the judgment unit 23 corrects the degree of negativity and positivity using the acquired information and predetermined correction data.

[0052] Furthermore, it is considered that the receiver's perception of what is said by the speaker (receiver) may differ depending on the speaker's (receiver's) age, nationality, sex, personality, etc. For this reason, it is also possible to correct the degree of negativity or positivity using attribute information of the speaker in the conversation, such as the speaker's age, nationality, sex, personality, etc. In this case, for example, the speaker's degree of negativity or positivity is corrected by the judgment unit 23 using the speaker's attribute information input to the information providing device 2 and pre-given correction data.

[0053] As described above, there are multiple types of information used to correct the negative and positive degrees of speakers. Here, the negative and positive degrees may be corrected by the judgment unit 23 using any one type of information, or the negative and positive degrees may be corrected by the judgment unit 23 using multiple types of information.

[0054] In the second embodiment, the determination unit 23 determines the corrected negative degree and positive degree as the negative degree and positive degree, respectively, corrected as described above. Using the determined negative degree and positive degree, the determination unit 23 determines (calculates) the harassment degree.

[0055] The information providing device 2 in the second embodiment has the same configuration as that of the information providing device 2 in the first embodiment, except for the configuration related to the correction of the determination unit 23 described above.

[0056] The information providing device 2 of the second embodiment has the same configuration as the first embodiment, and therefore can achieve the same effects as the first embodiment. In addition to the configuration of the first embodiment, the information providing device 2 of the second embodiment has a configuration for correcting the negative degree and the positive degree calculated using the number of negative words and the number of positive words included in the conversation. The correction is a correction of the negative degree and the positive degree taking into consideration that the receiver's feeling about the remarks differs depending on the situation of the conversation and the relationship between the speakers. The information providing device 2 can increase the accuracy of the harassment degree by determining (calculating) the harassment degree using the negative degree and the positive degree after such correction. In other words, the information providing device 2 can increase the reliability of the information on the harassment degree to be included in the harassment information.

[0057] <Third embodiment> A third embodiment of the present disclosure will be described below. In the description of the third embodiment, the same reference numerals are used to designate components that are the same as those in the harassment information providing system 1 of the first and second embodiments, and redundant descriptions of the common components will be omitted.

[0058] In the third embodiment, the information providing device 2 constituting the harassment information providing system 1 has, in addition to the configuration of the first or second embodiment, a configuration for determining the degree of harassment in consideration of the degree of tolerance (sensitivity) of each speaker (receiver) to harassing remarks. That is, as described above, possible factors that cause receivers to feel differently about remarks in a conversation include external factors such as the situation of the conversation and the relationship between the speakers. Furthermore, possible factors include internal factors of the speaker, such as each speaker's tolerance (sensitivity) to harassing remarks.

[0059] Therefore, in the third embodiment, the judgment unit 23 judges the degree of harassment by taking into consideration the degree of tolerance (sensitivity) of each speaker (receiver) to such harassment remarks. For example, the information providing device 2 acquires information (also referred to as harassment tolerance information) that indicates the degree of tolerance (sensitivity) of each speaker to harassment remarks. Then, the judgment unit 23 corrects the harassment degree calculated using the negative degree and the positive degree using the harassment tolerance information and correction data, and determines the corrected harassment degree as the harassment degree. The correction data here is relationship data between the degree of tolerance (sensitivity) to harassment remarks and the correction amount by which the harassment degree is corrected.

[0060] The following is an example of a method for acquiring harassment resistance information. For example, a questionnaire is administered to potential speakers of a conversation that is the subject of a harassment check to investigate their level of tolerance (sensitivity) to harassing remarks, and the responses to this questionnaire are used to investigate the level of tolerance (sensitivity) of each speaker (receiver) to harassing remarks. The survey results are input as harassment resistance information, and the information providing device 2 acquires the harassment resistance information.

[0061] The following method of acquiring harassment resistance information is also conceivable. For example, the information providing device 2 acquires voice data of a conversation in real time, and accepts input of information from the speaker of the conversation that indicates what the speaker feels is a harassing remark. After data in which the received information and the voice data of the conversation are associated is accumulated, the information providing device 2 calculates the degree of resistance (sensitivity) to harassment remarks for each speaker by analyzing the data, and thereby generates and acquires harassment resistance information. Note that instead of accepting input of information from the speaker of the conversation that indicates what the speaker feels is a harassing remark, for example, the information providing device 2 may acquire sensor output from a heart rate sensor of a wearable device worn by the speaker of the conversation or a captured image of the speaker's face. In this case, the information providing device 2 detects a change in the speaker's emotions using the information acquired in this way. Data in which this detection information is associated with the voice data is accumulated. The information providing device 2 analyzes the accumulated data to generate and acquire harassment resistance information.

[0062] It is believed that the above-mentioned harassment resistance information requires maintenance. In other words, it is believed that a person's level of tolerance (sensitivity) to harassing remarks changes depending on changes in the person's surrounding environment, changes in health condition, etc. For this reason, it is preferable that the information providing device 2 obtains updated harassment resistance information at a predetermined timing, for example.

[0063] The configuration of the information providing device 2 in the third embodiment is similar to that of the first or second embodiment, except for the configuration related to correction of the degree of harassment described above.

[0064] The information providing device 2 of the third embodiment has the same configuration as the first or second embodiment, and can therefore achieve the same effects as the first or second embodiment. Furthermore, the information providing device 2 of the third embodiment has a configuration that corrects the degree of harassment according to the tolerance (sensitivity) of the speaker (receiver) to harassing remarks. This allows the information providing device 2 of the third embodiment to determine the degree of harassment according to the degree of tolerance of the receiver of the conversation to harassing remarks. In other words, the information providing device 2 of the third embodiment can determine the degree of harassment that is appropriate to the actual situation.

[0065] <Other embodiments> The present disclosure is not limited to the first to third embodiments, and may be implemented in various ways. For example, the information providing device 2 may have a function for determining the timing of outputting harassment information in addition to the configurations of the first to third embodiments. That is, when information is presented to a person who made a harassment remark (harasser) during a conversation including the harassment remark, the warning may have an effect of suppressing harassment, or may have an adverse effect by upsetting the harasser. Therefore, the information providing device 2 monitors the state of the harasser using audio data and captured images during the conversation including the harassment remark. Then, the output unit 24 uses the monitored information and the harassment information output criteria to determine whether to output the harassment information during the conversation or after the conversation, and outputs the harassment information according to the result of the determination. The harassment information output criteria is information that represents a criterion for determining whether to output the harassment information during the conversation or after the conversation. This standard is established using the results of experiments and simulations that examine the relationship between the emotional state of the speaker (harasser) during a conversation and the effectiveness of presenting harassment information in reducing harassment.

[0066] By providing the information providing device 2 with a function for determining the timing of outputting such harassment information, the information providing device 2 can output the harassment information at a timing when the harassment information is likely to have an effect of suppressing harassment.

[0067] 6 to 8 show examples of display on a display device that displays harassment information, but the manner in which the harassment information is displayed is not limited, and may be, for example, a display manner as shown in FIG. 10. In the example shown in FIG. 10, the display of the harassment information is directed at the harasser. In the example of FIG. 10, a video (with audio) or still image showing the harasser during a conversation is displayed. In addition, the degree of harassment is displayed by a harassment gauge, and further, to clarify the danger of the degree of harassment, a warning level K, which indicates a level at which a warning is required, is displayed.

[0068] In addition, the harassment information providing device may have a configuration as shown in FIG. 11 as another embodiment. That is, the harassment information providing device 5 of the other embodiment is, for example, a computer device, and includes a transcription unit 51, a judgment unit 52, and an output unit 53, which are functional units realized by executing a computer program. The transcription unit 51 converts the speaker's remarks into text from the voice data in the conversation. The judgment unit 52 calculates the negative and positive degrees of the conversation by analyzing the text data representing the converted remarks, and judges the harassment degree of the conversation using the negative and positive degrees. The output unit 53 outputs harassment information including information representing the harassment degree. Examples of the configurations of the transcription unit 51, the judgment unit 52, and the output unit 53 are shown in the transcription unit 22, the judgment unit 23, and the output unit 24 in the first to third embodiments.

[0069] Next, an example of the operation of the harassment information providing device 5 will be described with reference to the flowchart in FIG. 12. For example, in a state where voice data of a conversation is acquired, the transcription unit 51 converts the speaker's remarks into text from the voice data of the conversation (step 201). Thereafter, the judgment unit 52 calculates the degree of negativity and positivity of the conversation by analyzing the text data representing the converted remarks. Furthermore, the judgment unit 52 judges the degree of harassment of the conversation using the calculated degree of negativity and degree of positivity (step 202). Thereafter, the output unit 53 outputs the harassment information (step 203). One of the destinations of the harassment information is, for example, a person who made a remark that raises concerns about harassment (the harasser).

[0070] The harassment information providing device 5 determines the degree of harassment taking into consideration not only the negative degree of a conversation but also the positive degree, and therefore can present convincing information that leads to the suppression of harassment to a person who has made a remark in a conversation that raises concerns about harassment. [Explanation of symbols]

[0071] 2 Information provision device 21 Acquisition Department 22,51 Transcription Department 23,52 Judgment section 24,53 Output section

Claims

1. a transcription unit that converts the speaker's utterances into text from the audio data in the conversation; a determination unit that calculates a negative degree and a positive degree of a conversation by analyzing text data representing a converted remark, and determines a harassment degree of the conversation using the negative degree and the positive degree; an output unit for outputting harassment information including information indicating the determined degree of harassment; A harassment information providing device comprising:

2. The harassment information providing device of claim 1, wherein the judgment unit extracts negative words, which are negative statements, and positive words, which are positive statements, from the text data of the conversation, calculates the degree of negativity using the number of negative words contained in the conversation, and calculates the degree of positivity using the number of positive words contained in the conversation, and judges the degree of harassment of the conversation using the degree of negativity and the degree of positivity.

3. The harassment information providing device of claim 2, wherein the judgment unit calculates the degree of negativity and positivity using information obtained from voice data in addition to information on the number of negative words and positive words contained in the conversation.

4. The harassment information providing device of claim 2, wherein the judgment unit calculates the degree of negativity and positivity using information on the number of negative words and positive words contained in the conversation, as well as at least one of information indicating the relationship between the speakers of the conversation and attribute information of each speaker.

5. 2. The harassment information providing device according to claim 1, wherein the judgment unit acquires harassment tolerance information indicating the degree of tolerance of each speaker to harassing remarks that may be considered harassment, and judges the degree of harassment using the harassment tolerance information.

6. By computer, Convert the speaker's remarks into text from the audio data of the conversation, Calculating the degree of negativity and positivity of the conversation by analyzing the text data representing the transcribed remarks, and determining the degree of harassment of the conversation using the degree of negativity and positivity; Output harassment information including information indicating the degree of harassment How to report harassment.

7. A process of converting the speaker's utterances into text from the voice data in the conversation; A process of calculating the degree of negativity and positivity of a conversation by analyzing text data representing the converted remarks and determining the degree of harassment of the conversation using the degree of negativity and positivity; A process for outputting harassment information including information indicating the degree of harassment; A computer program that causes a computer to execute the following:

Images