Hybrid AI-Human Intelligence Channel for Veracity Scoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The rapid spread of fake news on social media poses a significant challenge due to information overload and the lack of effective filtering tools, leading to damage in public confidence, reputations, and democratic processes, with existing solutions being inadequate in containing the spread in a timely manner.
Innovation Solution
A system utilizing a hybrid human expert network augmented by AI to create customizable real-time intelligence channels that filter out duplicates, identify relevant sources, and assign a Veracity Score to news content, enabling users to ban fake news sources and receive warnings, thereby reducing the spread of disinformation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If fact-checking processes are used to verify news content, then the accuracy of information is improved, but the time required to detect and respond to fake news increases significantly
Solution Approach 1:
The system performs preliminary actions by pre-identifying and flagging potentially fake news content using AI algorithms before it can spread widely. The hybrid human-expert network is prepared in advance to quickly verify flagged content, eliminating the need for time-consuming fact-checking of all news items. This preliminary filtering mechanism enables rapid response to disinformation while maintaining verification accuracy.
Solution Approach 2:
The patent introduces an intermediary AI-based detection system that sits between news content and human fact-checkers. This intermediary automatically identifies suspicious content and prioritizes it for human review, reducing the volume of content requiring manual verification. The intermediary layer enables faster detection by pre-processing and filtering content before it reaches the fact-checking stage.
2Quantity of substance
If traditional filtering methods are applied to news content, then information overload is reduced, but the effectiveness in containing disinformation spread is insufficient
Solution Approach 1:
The system employs an intermediary hybrid verification mechanism combining AI detection with human expert validation. This intermediary layer goes beyond traditional filtering by actively verifying suspicious content through a network of subject matter experts, ensuring that disinformation is identified and contained effectively while managing information flow.
Solution Approach 2:
The patent implements a feedback mechanism where the system continuously learns from verified fake news cases and updates its detection algorithms. The hybrid human-expert network provides feedback on AI detections, improving the system's ability to identify disinformation patterns. This feedback loop enhances both the filtering effectiveness and the containment of disinformation spread over time.
3Reliability
If organizational approval layers are implemented for public communications, then the reliability of responses is improved, but the turnaround time for defending against fake news increases
Solution Approach 1:
The system performs preliminary identification and verification of fake news content before organizational approval processes are triggered. By pre-flagging disinformation and preparing preliminary responses using the hybrid expert network, the system reduces the workload during crisis response, enabling faster deployment of verified counter-narratives while maintaining response quality through pre-established protocols.
Solution Approach 2:
The patent enables organizations to self-serve by providing them with access to the hybrid verification system and pre-prepared response templates. Organizations can independently verify fake news claims and access verified counter-information without going through lengthy external approval processes, significantly reducing response time while maintaining reliability through the system's verification mechanisms.
Data Source
AI summary
Techniques are provided for rating the veracity of content distributed via digital communications sources by creating an ontology and selecting keywords for a topic of the content, creating a customizable intelligence channel for the topic, and extracting from the customizable intelligence channel a first list of potential experts on the topic sorted by at least relevance and influence. The list of experts may be supplemented by mining trusted media sources to extract a second list of potential experts or witnesses on the topic. The first and second lists of potential experts are evaluated as a function of at least one of professionalism, reliability, proximity, experience, responsiveness, and lack of self-interest in the topic to identify a short list of experts. The content is provided to the short list of experts, who are polled about the veracity of the content to create a veracity score for delivery with the content.