Emakia ML Filter for Social Media Harassment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Social media users face harassment through unwanted content, and existing solutions, such as user reporting and content policing, are inadequate and often reactive, failing to create a safe environment for social media use.
Innovation Solution
The Emakia system employs machine learning classifiers to filter out harassing content at the receiver end, using Representational State Transfer (REST) API and Webhook to process incoming data from various social media platforms, separating unwanted content and allowing users to access filtered data while enabling automated reporting.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If machine learning classifiers are used to filter harassing content at the receiver end, then the safety and reliability of social media users is improved, but the device complexity and system complexity increases
Solution Approach 1:
The patent introduces an intermediary filtering system that operates between the social media platform and the user. This intermediary component, implemented as a machine learning classifier, automatically identifies and filters harassing content before it reaches the user's device. The system uses REST API and Webhook mechanisms to intercept and process incoming data, separating harmful content from legitimate content without requiring direct user intervention or modifying the core social media platform infrastructure.
2Productivity
If automated reporting is implemented, then the productivity and efficiency of harassment reporting is improved, but the device complexity and automation extent increases
Solution Approach 1:
The system implements self-service automation where the machine learning classifier automatically detects harassing content and triggers reporting mechanisms without requiring user intervention. When the classifier identifies harmful content, it automatically initiates the reporting process to appropriate authorities or platform administrators, eliminating the need for users to manually report each incident. This automated self-service approach significantly improves reporting efficiency while reducing the burden on users.
3Loss of time
If content filtering is applied at the receiver end, then the loss of time for users to access safe content is reduced, but the loss of information about filtered content increases
Solution Approach 1:
The system incorporates feedback mechanisms that inform users about the filtering actions taken. When content is filtered by the machine learning classifier, the system provides feedback to users explaining why the content was removed (e.g., flagged as harassment, spam, or inappropriate material). This feedback loop maintains user awareness of filtered content while ensuring users receive only safe and appropriate material, balancing time efficiency with information transparency.
Data Source
AI summary
Social media users are subject to harassment when unwanted offending content reaches them. Social media companies are reluctant to police content. The Emakia system provides a solution at the point where incoming real-time data are received. The Emakia system proposes to use sets of Machine Learning classifiers to filter text, images, audio, and video. Classifiers are trained with labeled data. After training, the model is used to screen the incoming real-time data. On the user mobile device, only approved content is displayed. The unwanted data are still available if the user desires to access them. The system provides multiple classifiers and customized models to the individual user. When harassment content is detected a report is sent to an entity who can help support the receiver.