Journalism + AI

Building Audience Analytics and Civic Integrity tools for Media Orgs.

Online trust in democratic institutions and newsrooms is eroding at an accelerated pace! One of the key challenges is the pace at which technology is being developed is much faster than the rate of adoption by journalists and newsrooms. With technology, it becomes possible to start to have a conversation about how to combat fake news, advance digital trust, and serve online audiences better!

The erosion of online trust has been a decades-long issue for US audiences, and is a worsening global trend!

We are actively working with newsrooms to help them seamlessly integrate technology into their editorial process and ensure their staff are equipped with the right skills to systematically improve the newsgathering and publishing process. We also help inform their digital marketing efforts and improve their understanding of online audiences.

Write to us to request a Demo of our Newsroom Analytics Platform

NYC Media Lab, 2021-23

Deutsche Welle (DW)

  • Hosted a 3.5-hour workshop for DW Akademie Staff and Global Partners on the Role of Media Development orgs. and Newsrooms in Combating Mis, Dis, and Malinformation (MDM) Operations
  • Invited talk and panel discussing the implications of generative AI for the newsroom (slides).
  • Invited short lab talk to DW Akademie and global partners discussing the detection and mitigation of influence operations and coordinated campaigns on social networks (slides).

Evaluating Media Sharing Across Social Networks

Our platform audits the impact of news articles and media orgs. in online information-sharing ecosystems such as social platforms. It is able to identify networks of accounts promoting content in a coordinated manner. This technology uses statistical methods to imbue accountability to the sharing of digital content, and understand the nature of how information spreads online. It supports the creation of public deliberation mechanisms for measuring the influence of accounts in the disjoint social ecosystem. The basis of this technology is the evaluation of content-agnostic sharing behavior in addition to heterogeneous datasets which allows it to generalize to different communities, languages, and countries without the need to explicitly provide local context. We develop and release these audits through an open-access platform that will allow users to review the spread of news on multiple social platforms.

Parrot Report

Screengrab of the Parrot Report website

Parrot is a joint project with The Sunday Times (UK), Ippen Media, and my team at SimPPL funded by the JournalismAI Fellowship at the London School or Economics. It is the precursor to the Arbiter Report project.

  • The Parrot platform demo with The Times (UK) was presented at the JournalismAI Festival 2022.
  • I’m attending the IPI World Congress–excited to meet the changemakers in journalism worldwide!
  • JournalismAI team Parrot comprising Ippen Digital (DE) and The Times / The Sunday Times (UK) is partnering with SimPPL for studying the spread of potentially manipulated narratives on Twitter from two well-known Russian state-backed media outlets.

Understanding the Underserved Interests of Visitors to a News Site

While media orgs. and newsrooms are able to understand their on-site visitors and on-platform audiences well, we developed a system that also tracks the audiences they aren’t successfully attracting to their platform, by utilising signals from social networks. We model audience preferences and conduct cohort analysis to ensure data on off-platform audiences is shared with the editorial and audience teams so that local newsrooms can better cater to new audiences and to the unserved interests of their existing audiences.

  • Whitepaper about our audience analytics technology is now available!
  • We are partnering with the oldest student newsroom, the Yale Daily News to launch an audience analytics pilot.
  • We ran a brief, successful pilot with the Vermont Digger to understand their off-platform audience interests (slides).