Methods and systems for protecting users of one or more social networks or video games from the risks of cyberbullying.

Abstract

The present invention relates to a method for protecting users of one or more social networks (30) or video games from the risk of cyberbuying using a connected device (7), the method comprising: performing semantic and contextual analysis on accessed or received digital content; employing artificial intelligence technology that implements multiple brains (C1-C5), each dedicated to a specific cyberbuying theme and operating simultaneously; and performing processing of the results of the semantic and contextual analysis to detect a cyberbuying situation or toxicity of all or part of the content. If a cyberbuying situation or toxicity of digital content is detected, the protection method according to the present invention writes (105) a set of data describing the cyberbuying to a blockchain (15) on the Web 3 and masks (104) the display of harmful digital content on the screen of the connected device (7).

Description

Technical Field

[0005]

[0001] The present invention relates to a method for protecting users of one or more social networks or video games from cyberbullying and the risk of harmful content. The present invention also relates to a processing system that employs this method.

[0002] The field of the present invention is more specifically the field of preventing cyberbullying and online toxicity on social networks.

Background Art

[0003] Since the emergence and widespread deployment of social networks around the world, the number of cases of cyberbullying and online toxicity has increased, and parents of children who have access to these social networks are powerless against the serious risks their children may face.

[0004] Software tools already exist for detecting harmful messages or cyberbullying messages in the flow of information passing through messaging systems and social networks.

[0005] U.S. Patent No. 11,138,237(B2) (IBM), “Social Media Toxicity Analysis,” discloses a method for analyzing the toxicity of social media content, including constructing a class model that represents the type and level of toxicity associated with the content. A first social media message is analyzed using a processor and memory to calculate a first toxicity score associated with the first social media message and assign the first social media message to a first class in the class model. A second social media message is analyzed using a processor and memory to calculate a second toxicity score associated with the second social media message and assign the second social media message to a second class in the class model. A toxicity trend is predicted based on the first toxicity score, the first class, the second toxicity score, and the second class. In response to a predicted increase in the toxicity trend, corrective actions are applied to reduce the toxicity trend.

[0006] U.S. Patent No. 10,915,789(B2) discloses a method and device for detecting offensive content on a portable electronic device by monitoring communications transmitted, received, or stored on a portable electronic device, including collecting content data, classifying the content data by calculating an alert score for the content data, the alert score corresponding to the detected offensive content, and sending an alert notification to a second portable electronic device to alert the detection of offensive content on a first portable electronic device.

[0007] U.S. Patent Application Publication 2019 / 0266242(A1) discloses a method for mitigating cyberbooling by intercepting and accurately analyzing the sentiment of social media messages before they are presented to the intended recipient's device. Based on the sentiment and sender category of the message, abusive messages may be sent to an administrator for review and appropriate action may be taken. As a result, such messages may be blocked from the intended recipient's device. The system may analyze the sentiment of a message using a machine learning approach, for example, employing a Naive Bayes algorithm. An inference model may be generated via machine learning. The inference model may generate metrics used when evaluating the sentiment of a message.

[0008] U.S. Patent Application Publication No. 2021 / 0042830(A1) discloses a digital multimedia system that uses a distributed communication system to transfer digital data. In some embodiments, the digital data includes digital assets, currencies, and instructions for executing financial transactions. In some embodiments, the digital multimedia system controls the storage and flow of digital data through designated nodes located in designated jurisdictions.

[0009] U.S. Patent Application Publication No. 2018 / 349502(A1) discloses a method and system for filtering content from a computing device's graphical display that may have objects associated with a content identifier. The method includes the computing device determining the weighted probability of each object based on instructions for selected filtering criteria, and determining the graphical elements of an application based on the weighted probabilities.

[0010] International Publication No. 2015 / 095597(A1) discloses a method and system that enables parents to view and track the activity of their children's smartphones, which may have one or more child software modules. Each child's smartphone can be installed with a module that can access and extract data relating to some of the smartphone's other software applications, including at least two of the following: text applications, social media applications, image applications that facilitate sending or receiving images, and web browser applications.

[0011] U.S. Patent Application Publication No. 2015 / 095014(A1) discloses a method for content selection and evaluation. A content classification platform processes content directed at or originating from a user to determine one or more elements within the content. A content evaluation platform calculates a content evaluation of the content by comparing one or more elements to at least one user-configurable dictionary, and then selects an action to process the content based on the content evaluation.

[0012] International Publication No. 2022 / 055546(A1) discloses a method, system, and product for filtering unwanted social multimedia content in real time. The system comprises a set of machine learning classifiers for filtering unwanted content on several media. The classifiers are trained on labeled data. After training, the model filters incoming data in real time, either on a server or on a mobile device.

[0013] However, it should be noted that the tools currently available are not effective or user-friendly for social network users and the parents of minors who access these networks.

[0014] The present invention aims to improve upon the shortcomings and limitations of existing tools by proposing a new method for the online processing of cyberbullying or harmful content that is more efficient and user-friendly than existing tools.

[0015] In this specification, the brain is understood as a computer model based on a simulation of the activity of a population of neurons that has or develops cognitive functions similar to those of a biological brain.

[0016] Blockchain can be understood as a large database with the specific characteristic of being shared simultaneously by all its users, all of whom are also owners of this register, and all of whom also have the ability to write data according to specific rules set by a highly cryptographically secure computer protocol, referring to the definition described in "Log of the joint information mission on blockchain and its uses: a sovereignty issue".

[0017] [https: / www2.assemblee-nationale.fr / static / 15 / commissions / CFinances / blockchain-synthese.pdf?v=1692355067]. [Overview of the Initiative]

[0018] The purpose of this is a method for protecting users of one or more social networks or video games who have a device connected to a communication network in order to access digital content from such one or more social networks or video games, or to receive digital content from one or more other users of such social networks or video games, from the risk of cyberbuying, - To employ artificial intelligence technology to perform semantic and contextual analysis on the accessed or received digital content, - Employing a database of predetermined situations, and performing processing of semantic and contextual analysis results in order to detect cyberbullying situations or toxicity in all or part of the content, -In the event that cyberbullying activity or toxicity of digital content is detected, this includes masking the display of harmful digital content on the screen of the connected device, It is to be implemented on a dedicated server suitable for connecting to the social network or video game via the said communication network, The method is characterized by further including writing a set of data describing the cyberbullying or toxic digital content to the blockchain when such cyberbullying or toxic digital content is detected, and by implementing multiple brains, each dedicated to a specific cyberbullying theme and operating simultaneously, each of which is intended to detect, after analysis, words, audio message transcripts, expressions, or emojis belonging to the associated theme.

[0019] Data describing cyberbullying or the toxicity of content may be output in the form of report logs.

[0020] The protection method according to the present invention may further include employing encryption of identification data entered by the user from a connected device and employing a distributed software tool on the Web3 to pre-check the identity of the user to be protected.

[0021] The protection method according to the present invention may further include the steps of accessing from a second connected device digital content received on a first connected device and detected as constituting cyberbullying or harmful content, and inputting from the connected device references to one or more social networks or video game networks for which protection is required.

[0022] When the protection method according to the present invention is applied to participation in an online video game, the semantic and contextual analysis step is further configured to detect a toxic situation in the video game.

[0023] According to another aspect of the present invention, a system for protecting users of one or more social networks or video games from the risk of cyberbuying, wherein the system employs a protection method according to the present invention and has a device connected to a communication network in order to access digital content from the one or more social networks or video games, or to receive digital content from one or more other users of the one or more social networks or video games, the system is -Means for employing artificial intelligence technology to perform semantic and contextual analysis on the accessed or received digital content, -Means for processing the results of semantic and contextual analysis in order to detect cyberbullying or toxicity in all or part of the content, by employing a database of predetermined situations. -Means for masking the display of harmful digital or cyberbulling content on the screen of a connected device if a cyberbulling situation or toxicity of digital content is detected, It is implemented on a dedicated server suitable for connecting to the video game social network via the communication network, and further comprises means for writing a set of data describing this server boiling to a blockchain, and the means for performing semantic and context analysis each has a plurality of brains dedicated to a specific server boiling theme and operating simultaneously, and each of the brains is intended to detect words, transcriptions of voice messages, expressions, or emojis belonging to the theme associated with it after analysis. A system is proposed characterized by this.

[0024] The protection system according to the present invention can be advantageously implemented in a dedicated server that communicates via one or more communication networks with, on the one hand, one or more platforms operating social network services and, on the other hand, a plurality of connected devices that access these social network services and require protection against server boiling.

[0025] It can also comprise means for sending a notification of the detection of a server boiling situation to a dedicated platform.

[0026] The processing method according to the present invention is advantageously executed in the form of an application for combating server boiling and online toxicity, and this application can be downloaded to any medium (mobile / tablet / pc / console).

[0027] With the help of artificial intelligence, this method masks the toxicity on the social network added by the user and the offensive messages that the user may receive. This analyzes the conversation semantically and contextually in order to obtain a result that matches the user's wishes.

[0028] If a user encounters an actual cyberbullying incident, an application implementing the method according to the present invention will authenticate the user's data to assist the user in legal proceedings and establish a file for filing a complaint with the competent authority. The application will also help victims find organizations specializing in combating online toxicity.

[0029] This application is also suitable for online gamers where toxicity is a recurring problem, as it integrates with the platform the online gamer is connected to, as well as the video game that authenticates the solution's operation.

[0030] This application also allows for notifications to be sent to the user and their designated guardian if they send inappropriate contextual elements that the recipient may consider harmful.

[0031] This application is built using software tools available in the AWS universe, following the architecture described below.

[0032] The processing method according to the present invention is - An application fully designed for deployment on Web3 across multiple media and social networks, -In contrast to the usual practice of using a single brain, AI employs multiple brains, each assigned to detect specific types of cyberbuying content, - It will adopt support for automatic output of report logs and writing of all types of content reports (images / text / videos / files) to the blockchain. [Brief explanation of the drawing]

[0033] [Figure 1] This is a block diagram of an exemplary embodiment of a system for protecting against cyberboiling according to the present invention. [Figure 2]The main steps of an exemplary embodiment of the protection method according to the present invention are illustrated below. [Figure 3] An example of a software architecture used in a protection system according to the present invention is provided. [Figure 4] This describes an example of a user interface that provides protected users with an overview of the actions taken to detect cyberbullying and harmful content. [Figure 5] This describes an example of a user interface that provides supervisory users with an overview of the actions taken to detect situations in which one or more family members, particularly minors, may have been exposed to cyberbullying and harmful content. [Figure 6] This describes an example of a user interface that provides blocking and reporting capabilities to protected users in response to the detection of cyberbullying or harmful content. [Figure 7] This describes an example of a user interface that provides video game users with a dashboard of detection operations performed by the protection method according to the present invention. [Modes for carrying out the invention]

[0034] Referring to Figures 1 and 2, the system 1 for protecting against cyberbullying according to the present invention is employed on a server 13 to which a database 14 is associated and connected via a common network to a digital cloud 2 (commonly referred to as the cloud) that links a set of social networking platforms 5 (or social media) including, but not limited to, Discord®, Tweeter®, Instagram®, TikTok®, or Facebook®, with a video game platform 4 such as the Steam® platform.

[0035] The method for protecting against cyberbulling according to the present invention is used in a set of digital communication devices connected to Cloud 2, such as smartphones 6-9, 12 or tablet 10, on which a dedicated application is installed. Some of these smartphones 8, 9 can connect to the resources of Cloud 2 via router 11.

[0036] The task of System 1 for protection against cyberbullying according to the present invention is to protect users of smartphones or tablets 6-10 equipped with a protection application from malicious acts by individuals or organizations from a server 5 connected to the cloud 2 that attempts to harm these users via one or more social networks or one or more video game platforms.

[0037] Now, with further reference to Figure 2, an embodiment of the method for protecting against cyberbullying according to the present invention will be described. It is assumed that the dedicated application is downloaded by the user of a smartphone 7 from a protection server 1 that implements the protection method according to the present invention.

[0038] This application, which provides a graphical interface on the screen of the user's smartphone 7, or on the screen of a smartphone 12 belonging to the parent or family member of a minor user, requests access 100 to the user's account on a social network, such as the Discord network, from the server 30 that manages the network.

[0039] When the server 13 managing the protection method gains access to the protected user's Discord social network account, semantic and functional analysis 101 of the content received on this user account is performed permanently or periodically at a predetermined frequency. This analysis is performed directly on the protection server before the analyzed content can be viewed on the user's smartphone 7.

[0040] This analysis 101 employs a set of artificial brains C1, C2, C3, C4, and C5, each assigned to a specific category of cyberbullying or toxicity.

[0041] Artificial brains are understood as computer models that possess or develop cognitive abilities similar to those of a biological brain, based on simulations of the activity of a population of neurons.

[0042] If a malicious server 5 sends a cyberbullying content message or a message containing harmful content to a protected user, one of brains C1-C5 will, after analysis, detect a word, audio message transcript, expression, or emoji belonging to one of the categories associated with each of these brains (103).

[0043] The protection server 13 then masks the message detected as constituting cyberbullying or harmful content (104) so ​​that the protected user cannot access the message or content on the screen of their smartphone 7.

[0044] This is followed by step 105, which involves writing data describing this detection of cyberbulling or toxicity to a blockchain 15 accessible by the protection server 13 via the cloud 2, and step 106, which involves outputting a report log 107 intended to be sent to the administrative or public agency responsible for combating cyberbulling.

[0045] The detected data may include, for example, the content of the intercepted message, the date, time, and (if technically possible) the source of the message.

[0046] In a configuration of the present invention specifically designed to protect underage users under parental supervision, several options can be provided for employing the protection method according to the present invention.

[0047] The protection method according to the present invention can be constructed from, but is not limited to, the software architecture 20 shown in Figure 3. Several basic modules of this software architecture can be obtained from AWS (Amazon Web Services) software offerings.

[0048] The software architecture 20 employs an SSL certificate management system 23 and a Domain Name System DNS 21 associated with a Content Delivery Network (CDN) for managing and hosting websites 25 that implement the protection method according to the present invention.

[0049] This website is built on a web framework 27 that drives a consent management module 24 and a personal identification and access management module 29.

[0050] The core of the software architecture 20 includes, for example, an image container module 34 built from the Python GraphQL framework and supplied by a Docker image registry 28.

[0051] This container 34 drives a payment service module 35, an image recognition module 30 controlled by an identification and personal access management module 29, a natural language processing (NLP) inference engine 36, a module 31 for managing identification and access to protected social networks, and a NoSQL database 37.

[0052] Module 31 for managing identification and access to social networks controls a serverless computing module 32 that directly addresses each of the social networks 21 (e.g., WhatsApp), 33 (e.g., Messenger) covered by the protection method according to the present invention via an authentication token. These social networks may include, but are not limited to, Discord, Twitter, Instagram, TikTok, or Facebook, as illustrated in Figure 1.

[0053] The natural language processing NLP inference engine 36 drives an NLP machine learning module 41 and an NLP personal learning module 42 that can address a GPU (Graphical Processing Unit) learning module 44 on demand, the GPU learning module 44 can be started by a time programming module 45.

[0054] The NoSQL database 37 is supplied to the data integration engine 38, which in turn supplies the data warehouse 39, equipped with a key performance indicator module 40, and the immutable database or blockchain 15.

[0055] Referring to Figures 4 to 7, several practical examples of graphical user interfaces generated on the screen of a protected user or supervisor by implementing the protection method according to the present invention will be described.

[0056] The user interface 50, represented by Figure 4, corresponds to what a protected user can see on their smartphone screen, and includes, for example, a circular graphic overview 51 showing the percentage of harmful messages 54 and blocked messages 55. Protected users can access overview data for each of the social networks 52 they subscribe to, and a list 53 of their most recent blocked messages.

[0057] Referring to Figure 5, a supervisor, for example, a parent or family member of legal age, can view the user interface 60, on which the supervisor can examine an overview 61 of the percentage of harmful messages 64 and blocked messages 65 for each of the supervised and protected users 62, 63.

[0058] Whenever cyberbullying or harmful content is detected, the protected user or their supervisor can view the “Protect Me” user interface 70 on their smartphone, which provides a “Block” action 71 that enables blocking the user sending the message, and a “Pharos Report” action 72 that enables constructing legal files and outputting reports based on detection information recorded by the protection method according to the present invention.

[0059] Naturally, the present invention is not limited to the embodiments described herein, and many other embodiments can be envisioned without departing from the scope of the invention. In particular, the number of artificial brains employed in the protection method according to the present invention is not limited to those described above. In addition, the structure of the detection data written to the blockchain can be of any type, depending on the effectiveness of the semantic and contextual analysis techniques employed. The same applies to the structure and actual content of the reporting logs.

Claims

1. A method for protecting a user of one or more social networks (3) or video games (4) from the risk of cyberbuying, wherein the user has devices (6-10) connected to a communication network (2) in order to access digital content from the one or more social networks (3) or video games (4), or to receive digital content from one or more other users of the one or more social networks (3) or video games (4), and the method is - Using artificial intelligence technology to perform semantic and contextual analysis (101) on the accessed or received digital content, - To employ a database of predetermined situations and perform the processing of the results of the semantic and contextual analysis (103) in order to detect cyberbullying situations or toxicity of all or part of the content, - If a cyberbullying situation or toxicity of digital content is detected, the display of harmful digital content on the screen of the connected device (6-10) is masked (104), It is implemented on a dedicated server suitable for connecting to the social network or video game via the aforementioned communication network (2), A method characterized by further comprising (105) writing a set of data describing cyberbulling or a toxic situation of digital content to a blockchain (15) if such cyberbulling is detected, and the semantic and contextual analysis implements a plurality of brains (C1-C5) that are dedicated to a specific cyberbulling theme and operate simultaneously, each of which is intended to detect, after analysis, words, audio message transcripts, expressions, or emojis belonging to the theme associated with it (103).

2. The method according to claim 1, characterized in that the data describing cyberbullying or the toxicity of the content is output in the form of a report log (107) (106).

3. The protection method according to claim 1 or 2, further comprising a pre-check of the user's identity to be protected by employing encryption of the identification data entered by the user from the connected device.

4. A protection method according to any one of claims 1 to 3, characterized by employing a distributed software tool on Web 3.

5. The protection method according to any one of claims 1 to 4, further comprising the step of accessing from a second connected device (12) digital content received on a first connected device (6 to 10) and detected as constituting cyberbullying or harmful content.

6. The protection method according to any one of claims 1 to 5, further comprising the step of inputting a reference to one or more social networks (3) or video game platforms (4) that require protection from the connected devices (6 to 10).

7. A protection method according to any one of claims 1 to 6, applicable to participation in an online video game (4), wherein the semantic and contextual analysis and processing steps are further configured to detect cheating activity in the video game.

8. A system (1) for protecting users of one or more social networks (3) or video games (4) from the risk of cyberbullying, wherein the user has devices (6-10) connected to a communication network (2) for accessing digital content from the one or more social networks (3) or video games (4), or for receiving digital content from one or more other users of the one or more social networks or video games, and employs the protection method described in any one of claims 1 to 7, and the system - Means (13) for performing semantic and contextual analysis on the accessed or received digital content using artificial intelligence technology, - A database (14) of predetermined situations is used to detect cyberbullying situations or toxicity of all or part of the content, and means (13) for processing the results of semantic and contextual analysis, - If cyberbullying or toxicity of digital content is detected, -Means for masking the display of harmful digital content on the screens of the connected devices (6-10), The system (1) is characterized by being implemented on a dedicated server suitable for connecting to the video game social network via the aforementioned communication network, further comprising means for writing a set of data describing the cyberbuling to a blockchain (15), and means for performing semantic and contextual analysis comprising a plurality of brains (C1 to C5) that are dedicated to a specific cyberbuling theme and operate simultaneously, each of which brains (C1 to C5) is intended to detect, after analysis, words, audio message transcriptions, expressions, or emojis belonging to the associated theme (103).

9. The protection system (1) according to claim 8, characterized in that it is implemented in a dedicated server that communicates via one or more communication networks (2) with one or more platforms operating social network services (3) on the one hand, and with multiple connected devices (6-10, 12) that access these social network services and require protection against cyberbullying.

10. The protection system (1) according to claim 9, further comprising means for transmitting a notification of the detection of a cyberbulling situation to a dedicated platform.

Images