Method and system for protecting a user of one or more social networks or video games from the risks of cyberbullying and toxic content

Abstract

The invention relates to a method for protecting a user of one or more social networks (3) or video games (4) from the risks of cyberbullying, wherein this user has a device (6-10) connected to a communication network (2) for accessing digital content originating from the one or more social networks (3) or video games (4), or for receiving digital content originating from one or more other users of the one or more social networks (3) or video games (4), and wherein the method comprises steps of performing a semantic and contextual analysis, of processing the results of the semantic and contextual analysis in order to detect a situation of cyberbullying or toxicity, and, in the event of detection of a situation of cyberbullying or toxicity in digital content, displaying the toxic digital content thus detected on the screen of the connected device. This method further comprises, upstream of the semantic and contextual analysis step, a step of collecting (data mining) digital content from a data stream received from a social network, and a step of recording, in a blockchain, a set of descriptive data describing this cyberbullying or toxicity.

Description

Method and system for protecting a user of one or more social networks or video games against the risks of cyberbullying and toxic content FIELD OF THE INVENTION

[0001] The present invention relates to a method for protecting a user of one or more social networks or video games against the risks of cyberbullying and toxic content. It also relates to a processing system implementing this method.

[0002] The field of the invention is more particularly that of the prevention and validity of proof of cyberbullying and online toxicity on social networks. TECHNOLOGICAL BACKGROUND

[0003] Since the emergence and massive deployment of social networks worldwide, an increasing number of situations of cyberbullying and online toxicity have been observed and parents of children with access to these social networks are helpless in the face of the serious risks that their children may run.

[0004] There are already software tools that aim to detect toxic or cyberbullying messages in the flow of information transmitted via messaging and social networks.

[0005] US 11,138,237 B2 (IBM) discloses a method for analyzing toxicity of social media content, including constructing a class model that is representative of a 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 of 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 of the class model. Based on the first toxicity score, the first class, the second toxicity score, and the second class, a toxicity trend is predicted.In response to the predicted increase in toxicity trend, a corrective action is applied to reduce the toxicity trend.

[0006] US 10,915,789 B2 discloses a method and apparatus for detecting offensive content on a portable electronic device by monitoring communications sent, received, or stored on the portable electronic device, and wherein the monitoring comprises collecting content data, classifying content data by calculating an alert score for content data wherein an alert score corresponds to detected offensive content, and sending an alert notification to a second portable electronic device to alert the detection of offensive content on the first portable electronic device.

[0007] US 2019 / 0266242 A1 discloses a method for mitigating cyberbullying by accurately intercepting and analyzing the sentiment of social media messages before presenting them to the intended recipient device. Based on the sentiment of the messages and the sender category, abusive messages may be sent to an administrator for review and appropriate action. As a result, such messages may be blocked from the intended recipient device. The system may analyze the sentiment of the messages using a machine learning approach, for example, using a Naive Bayes algorithm. An inference model may be generated via machine learning. The inference model may produce a metric that is used when evaluating the sentiment of the message.

[0008] US 2021 / 0042830 A1 discloses a digital media system that uses a decentralized communication system to transfer digital data. In some embodiments, the digital data includes digital property, currency, and instructions for executing financial transactions. In some embodiments, the digital media system controls the storage and flow of digital data through specified nodes located in specified jurisdictions.

[0009] US 2018 / 0349502 Al discloses a method for filtering content from a graphical display of a computing device. The content may have an object associated with content identifiers. The method further includes determining, by the computing device and based on an indication of selected filter criteria, a respective weighted probability for the object, and determining, by the computing device, a graphical element for the application based on the weighted probability.

[0010] WO 2015 / 095597 Al discloses a method for parents to view and track their children's smartphone activities. The module can be installed on each child's smartphone. The module can access and extract data regarding several other software applications on the smartphone, including at least two of the following: an SMS application, a social media application, an image application that facilitates the transmission or reception of images, and a web browser application. The module can also send the extracted data to an analytics server.

[0011] Document US 2015 / 0095014 Al discloses a content classification platform processes content directed to or from a user to determine one or more elements within the content. The content rating platform calculates a content rating for the content by comparing the one or more elements to at least one user-configurable dictionary, and then selects an operation to process the content based on the content rating.

[0012] WO 2022 / 055546 Al discloses a method for filtering unwanted social media content in real time. The system includes multiple sets of machine learning classifiers to filter unwanted content across any media type, including but not limited to text, images, audio, and video. Classifiers are trained with labeled data. After being trained, the models filter incoming data in real time either on a server or on a mobile device. A user application is executed, which displays only approved content on the main screen of the user application device. Unwanted data is still available if the user wishes to access it. The classifiers are trained with labeled data and with input parameters in addition to the labeled data.Custom models are trained with individual user data and transfer learning models. When unwanted content is detected, a report is sent to an entity that can provide assistance to the recipient for its handling.

[0013] Document US 2023 / 385975 A1 discloses a method for monitoring mobile communication and generating alerts associated with targeted content. The mobile communication device of a monitored user transmits an incoming or outgoing communication to the mobile communication device of a monitoring user. The mobile communication device of the monitored user or the monitoring user generates an alert if it is determined that the incoming or outgoing communication contains targeted content, and an alert is provided on the mobile communication device of the monitoring user.

[0014] However, these current methods of filtering and protecting against cyberbullying require the ability to interact with Application Programming Interfaces (APIs) associated with each of the social networks or messaging services installed on smartphones, which can make their effective deployment difficult.

[0015] The aim of the present invention is to overcome these drawbacks and limitations by proposing a new method of online protection against cyberbullying or toxic content, which presents a better level of efficiency and ergonomics than existing tools.

[0016] This objective is achieved with a method for protecting a user of one or more social networks or video games against the risk of cyberbullying, this user having equipment connected to a communication network to access digital content from said social network(s) or video game(s) or to receive digital content from one or more other users of said social network(s) or video game(s), comprising the following steps: a semantic and contextual analysis of said digital content accessed or received, implementing artificial intelligence techniques, processing of the results of semantic and contextual analysis to detect a situation of cyberbullying or toxicity of all or part of said content, implementing a database of predetermined situations, in the event of detection of a situation of cyberbullying or toxicity of digital content,a display of the toxic digital content detected on the screen of the connected equipment, and an entry in a blockchain of a set of descriptive data of this cyberbullying or this toxicity, said steps of semantic and conceptual analysis, processing and entry in a blockchain being carried out by a protection server connected to the communication network.,

[0017] According to the invention, the protection method further comprises, upstream of the semantic and contextual analysis step, a step for collecting said digital content received from said network, which is carried out by said protection server and comprises as many digital content collection processes as there are social networks to which the equipment is connected,

[0018] this protection server being able to access each of the said social networks and video games using access identifiers previously provided by the said user and / or their guardian,

[0019] This semantic and contextual analysis step implements a plurality of artificial intelligence brains, each dedicated to a particular theme of cyberbullying or toxicity and operating simultaneously.

[0020] Descriptive data of content toxicity can be advantageously edited in the form of a reporting report.

[0021] The protection method according to the invention may further comprise a prior check of the identity of the user to be protected, implementing encryption of the identity data entered by this user from the connected equipment.

[0022] This protection process can implement decentralized software tools on Web3.

[0023] It may further include a step for accessing, from a second connected device, digital content received and detected as cyberbullying or toxic on a first connected device, and a step for entering, from the connected device, a selection of one or more social networks or video game platforms on which protection is sought.

[0024] When applied to participation in an online video game, the semantic and contextual analysis and processing steps can further be configured to detect a toxicity situation in that video game.

[0025] According to another aspect of the invention, a system is proposed for protecting a user of one or more social networks or video games against risks of cyberbullying, implementing the protection method according to the invention, this user having equipment connected to a communication network to access digital content from said social network(s) or video game(s) or to receive digital content from one or more other users of said social network(s) or video game(s), this system comprising means for collecting said digital content from a data stream received in said connected equipment, means for carrying out a semantic and contextual analysis of said accessed or received digital content, implementing artificial intelligence techniques,means for processing results of semantic and contextual analysis to detect a situation of cyberbullying or toxicity of all or part of said content, implementing a database of predetermined situations, means for displaying, in the event of detection of a situation of cyberbullying or toxicity of digital content, the toxic digital content thus detected on the screen of the connected equipment,

[0026] According to the invention, this protection method is implemented within a dedicated protection server communicating via one or more communication networks, on the one hand, with one or more platforms operating social network services and, on the other hand, with a plurality of connected devices accessing these social network services and for which protection against cyberbullying is sought,

[0027] This protection server comprising said means of collection, semantic and conceptual analysis, and processing, and means for entering into a blockchain a set of descriptive data of a situation of cyberbullying or toxicity, and being able to directly access each of said social networks and video games by access identifiers previously provided by said user and / or his guardian.

[0028] The means of semantic and contextual analysis can advantageously include a plurality of artificial intelligence brains, each dedicated to a particular cyberbullying theme and operating simultaneously.

[0029] The protection system according to the invention may also advantageously comprise means for transmitting information to a dedicated platform reporting a situation of cyberbullying or a toxic message.

[0030] The treatment method according to the invention is advantageously implemented in the form of an application for combating cyberbullying and online toxicity, which can be downloaded onto any device (mobile / tablet / PC / console).

[0031] Using artificial intelligence, the process analyzes public and private social media feeds added by its users and the hate they may receive. It semantically and contextually analyzes conversations to obtain results that match the user's wishes.

[0032] If a user faces real cyberbullying problems, the application implementing the method according to the invention supports them in their legal procedures and certifies their data to establish a file for filing a complaint with the competent authorities. The application is also there to help the victim find an association specialized in the fight against online toxicity.

[0033] The application is also suitable for online gamers for whom toxicity is recurrent by integrating itself on the platforms on which they are connected as well as on video games allowing the solution to act.

[0034] The app also allows users and their designated guardians to be notified if they emit inappropriate contextual elements that could be considered toxic by their receivers.

[0035] This application is deployed using software tools available in the AWS universe, according to the architecture below:

[0036] The processing method according to the invention implements: an application fully designed for deployment on the cloud and Web3, multi-support and multi-social networks, An AI implementing a plurality of brains each assigned to the detection of a particular typology of cyberbullying content, in contrast to the usual practice of using a single brain, Assistance in the automatic editing of immutable evidence files and registration of content detected as cyberbullying (image / text / video / file / VOIP) in a blockchain. DESCRIPTION OF FIGURES

[0037] This is a block diagram of an exemplary embodiment of a system for protecting against cyberbullying according to the invention; This illustrates the main steps of an exemplary embodiment of the protection method according to the invention; This illustrates an example of software architecture implemented in a protection system according to the invention; This illustrates the graphical interface principles implemented in the cyberbullying method according to the invention; This represents an example of a home user interface displayed on a smartphone of a user in which a cyberbullying protection application has been installed; This represents an example of a user interface for managing guardian and relative accounts; This represents an example of a user interface providing a user with functionality for adding a guardian;L'il represent an example of a user interface that prompts a user to connect to a social network and then informs them that a data analysis is being undertaken;L'il represent an example of a user interface that provides a summary of the analysis results as well as functionality for adding a social network and functionality for managing guardians, with toxic messages detected;L'il represent an example of a user interface that provides a summary of the analysis results, in a configuration where toxic messages are not detected;L'il represent an example of a user interface that provides a display of the number of toxic messages sent by senders and the content of these messages;L'il represent an example of a user interface configured to defend itself and file a complaint in the event of detection of a toxic message or cyberbullying;This represents an example of a user interface that provides a user with functionality to send evidence of cyberbullying.;

[0038] DETAILED DESCRIPTION OF AN EMBODIMENT

[0039] With reference to Figures 1 and 2, a system 1 for protection against cyberbullying according to the invention is implemented in a server 13 with which a database 14 is associated, and is connected via a shared network to a digital cloud 2 (usually referred to as the Cloud) connecting a set of social network platforms 5 (or social media) including for example, but not limited to, Discord®, X® (formerly Tweeter®), Instagram®, TikTok® or Facebook®, and video game platforms 4 such as the Steam® platform.

[0040] The method of protection against cyberbullying according to the invention is used in a set of digital communication devices connected to the Cloud 2, such as smartphones 6-9,12 or tablets 10, in which a dedicated application has been installed. Some of these smartphones 8,9 can be connected to the resources of the Cloud 2 via a router 11.

[0041] The mission of the cyberbullying protection system 1 according to the invention will be to protect users of smartphones or tablets 6-10 equipped with the protection application, against the malicious actions of people or organizations from servers 5 connected to the Cloud 2 and seeking to harm these users through one or more social networks or one or more video game platforms.

[0042] Protection system 1 is designed to send, in the event of detection of cyberbullying messages, messages and evidence of reporting to dedicated servers 36 operated by sovereign authorities such as the police or the gendarmerie.

[0043] We will now describe, with reference to the, an example of implementation and operation of the method of protection against cyberbullying according to the invention. It is assumed that an application dedicated to protection has been downloaded by a user of a smartphone 7 from a protection server 1 implementing the protection method according to the invention.

[0044] Once the dedicated application has been activated, the data streams entering the smartphone 7 from social network platforms to which the user is subscribed will be subjected to a data collection process 100 comprising data mining operations 130-134 each corresponding to a social network, for example Discord, X, Instagram, Tik-Tok or Meta / Facebook. These data collections on the different networks are made possible because the user and / or their guardian has previously provided the operator of the method for protection against cyberbullying according to the invention with the access identifiers for these networks.

[0045] The data collected by data mining is subject to a semantic and functional analysis of the content received on this user account, which will be carried out either permanently or periodically according to a predetermined frequency. This analysis is carried out directly at the protection server before the analyzed content can be viewed from the user's smartphone.

[0046] This 101 analysis uses a set of artificial brains C1, C2, C3, C4, C5, each assigned a particular category of cyberbullying or toxicity.

[0047] An artificial brain is a computer model, based on the simulation of the activity of populations of neurons, which possesses or develops cognitive faculties similar to biological brains.

[0048] If a malicious server 5 sends a cyberbullying or toxic message to the protected user, one of the brains C1-C5 will detect 103, after analysis, a word, a transcription of a voice message, an expression, an image, a video or an Emoji belonging to one of the categories associated with each of these brains.

[0049] The protection server 13 will then display 104 the contents of the messages detected as cyberbullying or toxic content.

[0050] This is followed by a step 105 of recording descriptive data of this detection of cyberbullying or toxicity in a blockchain 15 accessible by the protection server 13 via the Cloud 2 and a step 106 of editing a reporting report 107 intended to be transmitted to an administrative or official authority responsible for combating cyberbullying.

[0051] Detection data may include, for example, the content of the intercepted message, the date, time and origin (if technically possible) of that message.

[0052] In a configuration of the invention dedicated to the protection of minor users under the supervision of a parent or guardian, several options for implementing the protection method according to the invention can be provided.

[0053] The protection method according to the invention can be constructed, without limitation, from a software architecture 20 such as that represented by the. Several basic modules of this software architecture can come from the AWS (Amazon Web Services) software offering.

[0054] The software architecture 20 implements a DNS domain name system 21 which feeds data to a collaborative platform 26, a corporate website 22, an application DNS zone 23 and an administration DNS zone 24. The corporate website 22 of the protection tool controls a Corporate Web Framework 38 which receives information from a consent management unit 37 and a content management unit 45.

[0055] The application DNS zone 23 feeds data to an authentication platform 27, an application 28, a content delivery network CDN 29 and an email service 30. The application 28 controls a web application framework 39. The content delivery network 29 feeds a GraphQL API 40 which transfers data to a transactional database 41. This transactional database 41 receives from a message collector 31, for example an AWS® Lambda module, data extracted from information flows coming from social networks, for example from the X® network or the Meta® network. The administration DNS zone 24 controls an administration authentication platform 32, an administration website 33, and a content delivery CDN module 34 which feeds a GraphQL API 42.

[0056] The software architecture 20 further comprises a daily management module “EveryDay” 25 which feeds a messaging system 35 associated with an image analysis module 36 as a service, a batch inference module 43 and an LLM module 44 for Large Language Model.

[0057] The transactional database 41 which receives data from both the application 40 and administration 42 GraphQL APIs, feeds a data integration module 96 cooperating with a data warehouse 93, and transfers cyberbullying or toxicity detection data to an immutable database 48 or to a blockchain 15 with reference to the.

[0058] The data integration module 96 also receives data from a payment provider module 95, a support center module 116, an accounting module 97, and social network servers such as Meta®.

[0059] An analytics tracking library 46 produces tracking pixels 91 and feeds a customer engagement platform 47 which is connected to the email service module 30 and an analytics platform 92 which feeds a data warehouse which is connected to an administration framework 49 itself feeding a digital identity and authentication center 90 which issues AWS (for Amazon® Web Services) accounts 98 and feeds a CRM (for Customer Relationship Management) module 99.

[0060] Software modules are dedicated to payment and accounting functions and services, such as the Payment Provider module 95 associated with a Bank module 94, or the Accounting module 97.

[0061] A software subset of the software architecture 20 comprises a design system module 110, a human-machine interface design module 111 cooperating with frameworks 112, 113. The GraphQM application cooperates with a web application framework 114.

[0062] Another software subset is dedicated to archiving and testing, including a Github code repository module 119, a Github CI / CD actions module 120, an infrastructure as code framework 122, a web end-to-end testing module 123, a mobile end-to-end testing module 124 and a hybrid application development module 125 cooperating with iOS® and Android® operating systems 121 that receive push notifications 118 generated by the customer engagement platform 47 and feed an identity verification module 115.

[0063] The software architecture 20 further includes a social media management platform 117.

[0064] As a non-limiting example of software implementation of the method for protection against cyberbullying according to the invention which has just been described, the message collector 31 can be carried out by a software unit of the AWS Lambda type. This software unit is executed each time a user provides information on a social network, which allows high scalability regardless of the number of users.

[0065] The messages thus collected are then transmitted to the artificial intelligence of the protection system according to the invention, which implements the GraphQL API 40, for analysis and detection of toxic messages.

[0066] The detected toxic messages are stored in a Neon unit, marketed by AWS, which is a "serverless" PostgreSQL solution. This Neon unit constitutes the transactional database 41.

[0067] Toxic messages, along with various user information, are stored in a QLDB 48 database that forms the AWS Blockchain, ensuring data immutability.

[0068] The 93 data warehouse is implemented as a Snowflake unit, marketed by AWS, which provides the Single Point of Truth function. All analysis and detection data points to the Snowflake 93 data warehouse in order to have the most complete data set possible, which makes it possible, in particular, to deliver performance indices and contribute to marketing operations, and to be able to easily cross-reference data from various sources.

[0069] We will now describe, with reference to figures 4 to 12, several practical examples of steps for displaying graphical user interfaces on the screen of protected users or on that of a tutor or supervisor, by implementing the protection method according to the invention.

[0070] With reference to the, these graphical user interfaces comprise, after downloading and installing a software application associated with the protection method according to the invention, a step of displaying a home interface (I) which in particular allows access to an interface (Ia) for adding a relative of the user who will be able to supervise the implementation of the protection method and an interface (II) for adding a tutor who will be able to supervise the application.

[0071] Step (I) of displaying the home interface leads to a step (III) of displaying an interface for connecting to a social network account held by the user, with the possibility of accessing an interface (Va) for adding a new social network.

[0072] Connecting to a social network account will lead to a step (IV) of displaying an interface indicating the execution of an analysis of the data exchanged by the user in this social network, which will lead to steps (V, VI) of displaying interfaces presenting a summary of the analysis results, with the possibility of managing relatives and guardians intervening for the user.

[0073] From step (V) of displaying the results summary, the user or his / her guardian will be able to access a step (VII,VIII) of displaying the number of toxic messages and their respective contents. The user or his / her guardian can then access a step (IX) of displaying a defense operations interface leading to dedicated interfaces (X) for reporting a toxic message to a reporting platform and for issuing a complaint in electronic format directly on dedicated servers operated by sovereign authorities such as the police or the gendarmerie.

[0074] As non-limiting examples of design of the graphical user interfaces which have just been cited, with reference to the, a home interface 60 (I) comprises a menu 63 comprising a set of actions selectable by the user, including the action 61 of adding a relative, the selection of which leads to the display of a window 62 for entering an e-mail address of this relative.

[0075] With reference to the, an account management interface 260 (Ia) comprises a window 261 for selecting between “my guardians” and “my relatives”, a window 262 for entering an e-mail address of a guardian or a relative to whom an invitation will be sent by means of an action on the command icon 263. “Add a relative”.

[0076] If a user selects the action “Add a tutor” 71 in the action menu 73 of the home interface 70 shown in, an input window 72 is displayed (II) to allow the user to enter an email address of a tutor that he wishes to add.

[0077] In order for the protection method according to the invention to be able to undertake the operations of detecting toxic messages, an interface 80 is displayed (III) to invite the user to enter personal data for connecting to a social network, for example Instagram, as illustrated.

[0078] This connection to the social network has the consequence of initiating the operations of collection by data mining and analysis of the data thus collected (IV), which are announced to the user by displaying a notification interface 90.

[0079] When the plurality of artificial intelligence brains have completed their operations of analyzing and classifying the collected data, an interface 200 providing a summary of the analysis results is displayed (V), as illustrated in. This summary interface 200 comprises a window 201 providing quantitative data 203, 204 on the number of toxic messages thus detected and on their proportion with respect to the total number of messages received, a window 202 indicating which social network is concerned by the current implementation of the protection method according to the invention and providing a functionality 205 for adding another social network, and a window 206 dedicated to the management of the tutors of this user.

[0080] If the functionality 205 for adding another social network is activated by the user, a network selection interface 240 is displayed (Va), with a window 241 for choosing favorite social networks and a choice icon or button 242.

[0081] The analysis of collected messages can also lead to the display (VI) of a summary interface 210 mentioning in a result display window 211 that no message has yet been collected for a social network displayed in a window 212, as illustrated.

[0082] From the results summary interface 200, the user can request (VII) an interface display 230 listing the numbers 231, 232 of messages received respectively from certain senders 241, 242 on a previously chosen social network 243, for example the Instagram network. In response to a selection of a sender 241 by the user, the protection method according to the invention generates, with reference to the, an interface 240 provided to display data 247 on the number of toxic messages detected, and contents 244, 245 of these toxic messages.

[0083] This interface 240 for displaying toxic content also includes a “Defend myself” icon or button 246, the selection of which by the user will lead to a display (IX) of proposals for defense actions in an interface 220 represented in.

[0084] This “defend me” interface 220 includes a window 221 for reporting a toxic message to a reporting platform such as Pharos and a window 222 for filing an online complaint with a dedicated platform. The reporting window 221 includes a “Defend me” icon or button 223, the selection of which leads to an automatic transfer of a file containing data relating to the toxic or cyberbullying message that is the subject of this report, while the complaint filing window 222 includes an icon or button 224, the selection of which leads to the display (X) of an interface 250 for choosing an e-mail address and sending evidence comprising two icons or buttons for choosing the police 251 or the gendarmerie 252, a window 253 for entering an e-mail address and a button or icon 254 for sending evidence by e-mail, as illustrated.

[0085] Of course, the present invention is not limited to the embodiment that has just been described, and many other embodiments can be envisaged without departing from the scope of the present invention. In particular, the number of artificial brains implemented in the protection method according to the invention is not limited to what has just been described. Furthermore, the structure of the detection data intended to be recorded in the blockchain can be of any nature, depending on the effectiveness of the semantic and contextual analysis techniques that are implemented. The same applies to the structure and practical content of the reporting reports and graphical user interfaces, the form and arrangement of the information displayed of which can vary without departing from the scope of the invention.

Claims

Method for protecting a user of one or more social networks (3) or video games (4), against risks of cyberbullying, this user having equipment (6-10) connected to a communication network (2) to access digital content from said social network(s) (3) or video games (4) or to receive digital content from one or more other users of said social network(s) (3) or video games (4), comprising the following steps: a semantic and contextual analysis (101) of said digital content accessed or received, implementing artificial intelligence techniques, a processing (103) of the results of semantic and contextual analysis to detect a situation of cyberbullying or a toxicity of all or part of said content, implementing a database of predetermined situations, in the event of detection of a situation of cyberbullying or a toxicity of digital content,a display (104) of the toxic digital content detected on the screen of the connected equipment (6-10), and an entry (105) in a blockchain (15) of a set of descriptive data of this cyberbullying or this toxicity, said steps of semantic and conceptual analysis, processing and entry in a blockchain being carried out by a protection server (13) connected to the communication network (2), characterized in that it further comprises, upstream of the semantic and contextual analysis step, a step for collecting said digital content received from said network, which is carried out by said protection server (13) and comprises as many digital content collection processes as there are social networks to which the equipment is connected, said protection server (13) being able to access each of said social network(s) and video games by access identifiers previously provided by said user and / or his tutor,said semantic and contextual analysis step implementing a plurality of artificial intelligence brains (C1-C5) each dedicated to a particular theme of cyberbullying or toxicity and operating simultaneously., Method according to the preceding claim, characterized in that the descriptive data of a content toxicity are edited (106) in the form of a reporting report (107). Protection method according to any one of the preceding claims, characterized in that it further comprises a prior check of the identity of the user to be protected, implementing encryption of the identity data entered by this user from the connected equipment. Protection method according to any one of the preceding claims, characterized in that it implements decentralized software tools on Web3. Protection method according to any one of the preceding claims, characterized in that it further comprises a step for accessing from a second connected device (12) digital content received and detected as relating to cyberbullying or toxic on a first connected device (6-10). Protection method according to any one of the preceding claims, characterized in that it further comprises a step for entering from the connected equipment (6-10) a selection of one or more social networks (3) or video game platforms (4) on which protection is sought. Protection method according to any one of the preceding claims, applied to participation in an online video game (4), characterized in that the steps of semantic and contextual analysis and processing are further configured to detect a toxicity situation in this video game. System for protecting a user of one or more social networks (3) or video games (4), against risks of cyberbullying, implementing the protection method according to any one of the preceding claims, this user having equipment (6-10) connected to a communication network (2) to access digital content from said social network(s) (3) or video games (4) or to receive digital content from one or more other users of said social network(s) (3) or video games (4), this system comprising means for collecting said digital content from a data stream received in said connected equipment, means for carrying out a semantic and contextual analysis (101) of said accessed or received digital content, implementing artificial intelligence techniques,means for processing (103) results of semantic and contextual analysis to detect a situation of cyberbullying or toxicity of all or part of said contents, implementing a database of predetermined situations, means for displaying, in the event of detection of a situation of cyberbullying or toxicity of digital content, the toxic digital content thus detected on the screen of the connected equipment (6-10), characterized in that it is implemented within a dedicated protection server (13) communicating via one or more communication networks (2), on the one hand, with one or more operating platforms of social network services (3) and, on the other hand, with a plurality of connected equipment (6-10, 12) accessing these social network services and for which protection against cyberbullying is sought, said protection server (13) comprising said means of collection, semantic and conceptual analysis,and processing, and means (105) for entering into a blockchain (15) a set of descriptive data of a situation of cyberbullying or toxicity, and being able to directly access each of said social networks and video games by access identifiers previously provided by said user and / or his tutor., Protection system (1) according to the preceding claim, characterized in that the semantic and contextual analysis means comprise a plurality of artificial intelligence brains (C1-C5) each dedicated to a particular cyberbullying theme and operating simultaneously. Protection system (1) according to the preceding claim, characterized in that it further comprises means for transmitting to a dedicated platform information reporting a situation of cyberbullying or a toxic message.

Images