Today, our society is struggling with an unprecedented amount of falsehoods, hyperboles and half-truths that do harm to democracy, health, economy and national security. Fact-checking is a vital tool for defending against this onslaught. Despite the rise of fact-checking efforts globally, fact-checkers find themselves increasingly overwhelmed and find it difficult to reach some segments of the public with their messages.

This project team sought to leverage the power of data and computing to help make fact-checking and dissemination of fact-checks to the public more effective, scalable and sustainable. Building on the work of an affiliated Summer 2018 Data+ team, the project team designed an interface for data verification.

First, the team collected audio from newscasts and YouTube videos. These were then transcribed via Google Cloud Speech-to-Text-API and checked for check-worthy factual claims using ClaimBuster. Through the use of information retrieval and natural language processing techniques each claim was matched in a database of fact-checks.

Ultimately, the team plans to deliver a live fact-check stream that can be played over live broadcasts and YouTube videos. The current version has 65-80% accuracy, thus human input is needed for better curation. The team has introduced the tools to journalists who can use it to monitor and update claims in real time. The team did its first live test-run of the tool on the 2019 State of the Union speech. Team members hope to have the next generation of the pop-up fact-checking system, SQUASH, available for use for the 2020 U.S. elections.

Timing

Fall 2018 – Spring 2019  

Team Outputs

Automated Fact-Checking App. William Adair, Pankaj Agarwal, Jun Yang, Towqir Aziz, Mayuresh Kunjir, Yuhao Wen, Archana Ahlawat, Sherry Feng, Drew Learner, Matthew O'Boyle, Jason Wang, Connie Wu, Fengyu Xie, Liuyi Zhu, James T. Hamilton, Chengkai Li, Cong Yu. 2018.

Convergence Accelerator Phase 1 (RAISE): Credible Open Knowledge Network. Jun Yang. National Science Foundation. 2019. $999,870.

Developing Automated Tools for Political Fact-checking (poster by Matthew O’Boyle, Sherry Feng, presented at Bass Connections Showcase, Duke University, April 17, 2019; runner up, Bass Connections Poster Competition, Judges’ Selection)

Data and Technology for Fact-checking (poster by Archana Ahlawat, David Cheng, Fangge Deng, Lucas Fagan, Sherry Feng, Yuanhao Guan, Ethan Holland, Sherry Hu, Drew Learner, Matthew O'Boyle, Ali Soyupak, Caroline Wang, Jason Wang, Connie Wu, Harry Xie, Arthur Zhao, Liuyi Zhu, Towqir Aziz, Xinghao Cheng, Mayuresh Kunjir, Junbo Li, Yuhao Wen, Bill Adair, Pankaj K. Agarwal, Jun Yang, presented at Bass Connections Showcase, Duke University, April 17, 2019)

Challenges in Automated Pop-up Fact Checking (poster by Fangge Deng, Yuanhao Guan, Ali Soyupak, Jason Wang, Connie Wu, Arthur Zhao, Xinghao Cheng, Mayuresh Kunjir, Junbo Li, Bill Adair, Pankaj K. Agarwal, Jun Yang, presented at Bass Connections Showcase, Duke University, April 17, 2019)

Video

Automated Fact-checking App

This Team in the News

Dora Pekec Makes Every Minute Matter at Duke and Beyond

Number of Fact-checking Outlets Surges to 188 in More Than 60 Countries

Fact-checking the President in Real Time

Archana Ahlawat Emphasizes Improving Experiences of Underrepresented Groups in Young Trustee Campaign

Live Fact-checking of the State of the Union Address with Our FactStream App

Free App Provides Live Fact-Checking During Tonight's State of the Union Address

Real Facts in Real-time: Duke Team Aims to Develop First Live Fact-checking Tech by 2020 Election

US Team Working on Real-time Fact-checking App

Technology Near for Real-time TV Political Fact Checks

Real-Time Fact-Checking: There’s (Almost) An App For That

Duke Students Tackle Big Challenges in Automated Fact-checking

The Red Couch Experiments: Early Lessons in Pop-up Fact-checking

Bass Connections Team Pursues First-ever Automated Fact-checking App

See related team, Data and Technology for Fact-checking (2019-2020), and a related Data+ summer project, Data and Technology for Fact-checking (2018).