Containing disinformation spread using customizable intelligence channels
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.
1. A computer-implemented method of rating the veracity of content distributed via digital communications sources, comprising:
creating an ontology and selecting keywords for a topic of the content;
creating a customizable intelligence channel for the topic of the content and extracting from the customizable intelligence channel a first list of potential experts on the topic of the content;
mining trusted media sources for the topic of the content to extract a second list of potential experts on the topic of the content;
providing the first and second lists of potential experts on the topic of the content to a database;
rating and ranking the potential experts based on a combination of factors selected from professionalism, reliability, proximity, experience, responsiveness, and lack of self-interest in the topic of the content to identify a short list of experts, wherein proximity is a measure of trustworthiness identifying a closeness in space, time, or relationship of the potential expert to the topic of the content;
providing the content to the short list of experts for evaluation;
polling the short list of experts about the veracity of the content to create a veracity score;
delivering the veracity score; and
benchmarking the veracity scores to create a predictive fake news spread containment model and iterating to revise the model and overall performance of the model over time.
2. The method of claim 1 , further comprising creating a third list of potential experts and any local witnesses on the topic of the content based on at least one of a relationship and a proximity of the potential experts to a breaking story on the topic of the content and providing the third list of potential experts to the database to complement the polling.
3. The method of claim 1 , wherein delivering the veracity score comprises at least one of issuing a pre-populated press release, initiating a social media and press campaign including the veracity score and at least one of a warning and denial when the content is not completely true, issuing a quote from an expert from the short list of experts, issuing a quote from a local witness to the topic of the content, and issuing a quote from an influencer on the topic of the content and related reassuring metrics including information about trustworthiness of sources of the content.
4. The method of claim 1 , wherein a fake news warning is presented with the veracity score and content along with insights and metrics relating to the content.
5. The method of claim 4 , wherein the veracity score, content, fake news warning, insights and metrics relating to the content are delivered via an interactive interface enabling a user to select the types of sources by level of trust or proximity to the news content or user wherein the proximity of the type of source to the news content or user is a measure of trustworthiness identifying closeness in space, time, or relationship.
6. The method of claim 1 , wherein providing the first and second list of potential experts on the topic of the content to a database includes predetermining a list of experts to crowdsource veracity signals for a given topic of content.
7. The method of claim 1 , wherein mining trusted media sources for the topic of the content to extract a second list of potential experts on the topic of the content is performed upon the release of a news story.
8. The method of claim 1 , further comprising incentivizing experts to be accurate and prompt when consulted by compensating experts based on how accurate and prompt the experts' predictions are and creating a decentralized register including expert trust ratings.
9. The method of claim 1 , further comprising creating at least one customizable intelligence channel for the topic of the content relating to potential sources of fake news and the semantics of fake news content.
10. The method of claim 1 , further comprising creating a decision matrix to evaluate a breaking news story to decide whether the news story is a candidate for determining a veracity rating based on at least one of the nature of the breaking news story, a source of the news story, and whether a relevant population of experts readily exists.
11. The method of claim 1 , further comprising modifying the veracity score to reflect the behavior of additional experts and trusted sources in sharing and commenting upon a breaking news story.