DemLabs blog

How to automatically monitor hate speech on talk radio with AI

"Conservative Talk Radio Spreads Hate Speech Targeting Minorities, Religious Groups and the LGBT Community" - NHMC

Use AI to quickly monitor hate speech on talk shows and radio channels? Use AI.

The "Social Networks for Hate Speech" report analyzed the role that five prominent conservative talk radio shows play in spreading hate speech on social websites. It examined the guests, topics and content on The Rush Limbaugh Show, The Sean Hannity Show, The Glenn Beck Program, The Savage Nation, and The John and Ken Show.

This blog explains how we used AI to find mentions of certain words and phrases on a talk show program. The same approach also works to search for specific content in videos and podcasts.

How to monitor radio talk shows for hate speech

Listen. Analyze. Report.

You can find dozens of radio stations to listen to at Tune In. We chose this channel clip at random to demonstrate how the app is used, not because the audio clip contained hate speech. The sound recording was uploaded into this (currently) free DemLabs app and asked for it to scan for the word 'shooter'.

Try it out for yourself with any sound recording, video or podcast. Just upload the file and choose the word that you're interested in monitoring. The app will report where that word was said and provide you with a transcript and link to start playing the recording at that time the word was said.

Launch QuickTime Player from the Applications folder.  Drag and drop the audio file onto the QuickTime Player Dock icon or into the QuickTime app. Set the timer to where the word you were searching for was found and click the play button.

The app is being enhanced to transcribe broadcasts and conversations in real time. The app will also produce summaries, checklists and be able to search for phrases and sentiment, in addition to words. Learn more about the app and other use cases for here.

Dog whistle politics send coded messages

Real time monitoring of hate speech and disinfo is important

This network include Premiere Radio Networks, a wholly-owned subsidiary of Clear Channel Communications that nationally syndicates three of the five shows examined, and Fox News, whose media personalities appear most frequently on the shows. The study completed over 18 months found:

  • The social network of hate speech propagates a narrow set of priorities, such as messaging that often includes anti-immigration and anti-Islamic focuses.
  • In this highly insular network, immigrants are routinely referred to as “illegal” and law- breaking, while Islam is portrayed within the context of terrorism.
  • An overwhelming proportion of scheduled guests on these shows were white (89%) and male (81%), matching the race and gender of the hosts.
  • Political figures who appeared as guests were largely members of the Republican Party (93%) and/or affiliated with the Tea Party movement (89 percent).
  • The top five ideological positions of scheduled guests were determined to be anti- federalist (23.6%), pro-Republican (19.1%), pro-Tea Party Movement (12.1%), anti-Islamic (11.1%)and anti-immigration (8.5 percent).
  • Though 88 topics of interest were identified, immigration emerged as the most discussed topic, dominating nearly one-quarter of all topic segments. - NHMC

Free speech and Hate Speech

The First Amendment of the Constitution broadly protects most forms of speech. Just because a viewpoint is controversial does not mean that it should be quelled. The Supreme Court and other US courts have ruled that speech may be protected even if it is vulgar and offensive, subject to some narrow restrictions. This constrains the ability of broadcast stations and the FCC to restrain speech or respond to complaints.

The FCC retains enforcement powers in certain situations to which more limited protections extend. These generally arise in the context of obscene, indecent, or profane material. While these words are often used interchangeably, they have different meanings in the context of the First Amendment and communications law. The Supreme Court has developed a three-part test for determining when content is obscene:
- Would the average person find that the work as a whole appeals to the prurient interest, considering contemporary community standards?
- Does the work depict or describe sexual conduct as defined by state law in a patently offensive way?
- Does the work as a whole lack any serious literary, artistic, political, or scientific value?

Violations of these rules can lead to monetary penalties, license revocation, or the denial of a renewal application. The US Department of Justice also can pursue criminal charges in some cases, which can result in fines and imprisonment. - Justia

TakeAway: Monitor hate speech and disinformation on Talk Radio with AI.

DISCLAIMER: ALTHOUGH THE DATA FOUND IN THIS BLOG AND INFOGRAPHIC HAS BEEN PRODUCED AND PROCESSED FROM SOURCES BELIEVED TO BE RELIABLE, NO WARRANTY EXPRESSED OR IMPLIED CAN BE MADE REGARDING THE ACCURACY, COMPLETENESS, LEGALITY OR RELIABILITY OF ANY SUCH INFORMATION. THIS DISCLAIMER APPLIES TO ANY USES OF THE INFORMATION WHETHER ISOLATED OR AGGREGATE USES THEREOF.

Share this Newsletter:

Share on twitter
Share on linkedin
Share on facebook

DemLabs applies innovative technology and storytelling tools in service of democratic values. It lowers the barrier of funding for worthy candidates and non-profits by applying existing free/affordable solutions.