Seekr, an artificial intelligence company specializing in content evaluation, says it has received a new patent (No. 11,893,981) for developing a scoring system that identifies personal attacks in online audio content. Beginning with podcasts and extending to transcripts, closed captioning, audible books, and social media, Seekr AI contextually evaluates audio and generates what it calls a Civility Score that reflects the volume and severity of attacks within each piece of content.
Rob Clark, President and Chief Technology Officer at Seekr, says it is another way artificial intelligence is being used to solve a business challenge, which in this case, is how to accurately evaluate the brand suitability of audio programming in a market comprising hundreds of millions of audio episodes.
"This challenge could only have been solved by combining superior computing power with novel signal detection," Clark says. "The Civility Score synthesizes dozens of considerations into a single metric that opens the door to more brand-suitable opportunities and larger audiences."
Seekr formally introduced the Civility Score in January with an initial application as a brand safety and suitability solution for podcasts. After months of beta testing, Seekr produces a standard civility score that will give marketers a number on a 0 to 100 scale with a clear specific explanation of why a show or individual episodes achieved that score. Seekr will also allow subscribers to see what exactly in the discussion led to a specific civility score so they can assess how the show stacks up based on their specific risk tolerance. The platform also serves up similar shows, giving an ad buyer a sense of what else might be available for similar categories, topics or themes.
"The Civility Score establishes a new standard that goes far beyond current brand safety and suitability standards by recognizing that how a topic is discussed is as important as what is being discussed," Clark says.
Seekr says it has so far scored more than eight million minutes of audio across hundreds of thousands of podcasts within Edison Research’s ranking of the top 1,000 shows. It expects to have scores for ten-times as many podcasts by next year for the service it is calling Align.
コメント