Meedan is pleased to roll out Check Bot: a fully customizable bot designed for fact-checkers and journalists to run fact-checking tiplines on Whatsapp, Facebook Messenger, Twitter DM, and more.
The bot, which also works on other social messaging platforms, is designed to scale response through the COVID crisis and other topical content. Now being tested in Brazil, India and across Africa with COVID-19 content, our bot is an extension of Check, our collaborative platform for digital media verification and annotation, to boost the work of human journalists through AI-powered claims matching, smart menus, and the ability to source new content directly from users for fact-checking.
It has the following key features:
- A fully customizable interaction scenario , allowing fact-checking teams to spin up interactive menus of frequently-requested content to react to events rapidly
- Enables audience-submitted tips for queries not answered by the bot
- Ties directly to a team’s Check database, allowing them to send previously fact-checked reports
For our partner organizations, the bot is an essential communication channel to engage their audience, size the demand for specific topics, and distribute key information in close messaging networks. Through building this relations, it is also the opportunity to develop interactive monitoring and evaluation surveys with end-users to assess behavioural impacts of fact-checks.
Check Bot: a fully customizable bot designed for fact-checkers and journalists to run fact-checking tiplines on Whatsapp, Facebook messenger, Twitter DM, and more.
When designing this bot, we worked with our fact-checking partners to ensure the following principles were met:
- Audience development: Fact-checking teams have spent years cultivating trust with their audiences, and we wanted to make sure the bot supports this work. Teams can customize every interaction and content that represents their unique voice. They can continue to receive fact-check requests through the same tipline and communicate with their audiences directly.
- Ready for scale: We are testing the bot with teams that receive thousands of messages per week. The bot is designed to handle requests of many thousands more, automating frequently-requested content while overall improving the quality of tipline content submissions.
- Local content, local languages: This bot can provide more detailed content coverage for local audiences that complements the work of global health organizations and national health ministries. Fact-checking organizations can create custom resource responses for their end users that are context or language-specific, directly meeting a target population’s health information needs. Thanks to our partnership with the Localization Lab, we can also localize the content and interface to relevant languages.
Right now, we’re testing this bot with COVID-19 content and localized COVID-19 resources. Per our ongoing commitments to open source and the Open COVID Pledge, it is also open source. The goal is to link this up with Meedan’s COVID-19 Expert Database, developed by our Digital Health Lab. This database contains contextualized health information developed and vetted by public health experts that can be used as source material in fact-checks, removing some of the challenges of accessing verified expert content
The bot can be fully adapted for any kind of content, addressing audience queries related to politics, science and breaking news events. We are planning to add more messaging services such as Lime, Telegram, Viber, Instagram and many more.
Interested in a demo? Get in touch at email@example.com.
- Online conversations are heavily influenced by news coverage, like the 2022 Supreme Court decision on abortion. The relationship is less clear between big breaking news and specific increases in online misinformation.
- The tweets analyzed were a random sample qualitatively coded as “misinformation” or “not misinformation” by two qualitative coders trained in public health and internet studies.
- This method used Twitter’s historical search API
- The peak was a significant outlier compared to days before it using Grubbs' test for outliers for Chemical Abortion (p<0.2 for the decision; p<0.003 for the leak) and Herbal Abortion (p<0.001 for the decision and leak).
- All our searches were case insensitive and could match substrings; so, “revers” matches “reverse”, “reversal”, etc.