Webinar No. 1 presenting veraAI outcomes 

Welcome and introduction

After a quick introduction to the day's event and proceedings, also outlining the agenda and context, moderator Jochen Spangenberg (DW) gave the floor to project coordinator Akis (Symeon) Papadopoulos from CERTH-ITI. Akis gave a high level overview of veraAI, pointing to its approach, components and respective research challenges.

Denis Teyssou (AFP) presenting the verification plug-in

A core result of the project is the so-called verification plug-in. Started in the EU project InVID, further developed in WeVerify, it was and is still in the process of being enhanced further in veraAI. In other words: additional components were and are added to it, among them those dealing with synthetic media detection and analysis.  

The plug-in now has more than 150,000 regular users. It is available free of charge, optimised for use with the Chrome browser. To get your own version go to the Chrome store and install it as an add-on. 

Note: Denis did a live demo - hence no presentation is included below. You can view his contribution in the recording (it starts at minute 18 in the recording that you find at the bottom of this article).

Danae Tsabouraki (ATC) presenting Truly Media enhancements

Truly Media also has been around for some time, having its foundations in EU co-funded research. It is a commercial platform for collaborative versification of digital content, developed and operated in a public-private partnership by ATC and DW. 

In the course of veraAI, Truly Media was enhanced, too, with components developed in the project. Most are still in the trial or deployment stage. 

In her presentation, Danae gave an overview of the work done and how this can help and support Truly Media customers in their verification workflows.

The Database of Known Fakes (Andrey Tagarev, Ontotext)

The idea behind the so-called Database of Known Fakes (DBKF in short) is manifold:

• host a collection of already verified / debunked items,
• make this searchable, also across multiple languages,
• reduce the duplication of work in fact-checking,
• allow for comparisons across languages,
• cluster information around topics,
• and much more.

Andrey Tagarev (Ontotext, now trading as Graphwise) showcased features, components and development status of the DBKF. He furthermore invited webinar attendees to test and trial the database themselves.

Discussion and break

This first part was followed by a discussion of moderator Jochen Spangenberg with the three panelists. The audience took part actively. Main topics: 

• challenges to date and what can be expected in the future,
• going more into depth concerning what had been presented beforehand. 

For details of the discussion see the recording (included at the end of the article).

Then followed a 15 minute break before we resumed the webinar to enter the second and final round of the event. 

Part 2: All about Coordinated Inauthentic Behaviour (CIB)

Part 2 of the webinar was moderated by Ines Gentil of EUDL. It included three presentations - they all focussed on CIB, or in full: Coordinated Inauthentic Behaviour. 

Presenters were, in this order:

• Ana Romero (EUDL)
• Nicola Righetti (Uni Urbino)
• Kamilla Koronska (Uni Amsterdam)

Below, their presentations are included. For details of the different approaches and findings we again recommend viewing the video that also includes these presentations (it all starts at 1:34'26'').

Recap of the afternoon

At the end of a busy and information-packed webinar, I (the author of this article) tried to wrap up the day. Not an easy task. I attempt to do so below again - using the notes I took in the course of our session:

"I hope we managed to give you both – an idea of what the vera.ai project and work is about and some of what we have achieved so far.

In particular, we hope there was something useful in it for you, who came from all over the world, which we appreciate very much.

In the first session, we gave you insights into some of the veraAI tools and services that are there to support fact-checkers, journalists, disinformation detectors and others active in the field. Much of our work and results is available to all – so get your hands dirty if you like. And let us know your experience and give us your feedback. We would appreciate this very much.

In the second part, we then took a deep dive into the way disinformation spreads – doing harm and undermining our democracies, eroding trust and much more. 

While all this requires utmost attention and needs to be dealt with, it is also of interest from a scientific perspective. Additionally, it also provide for great storytelling.

In the course of the afternoon we also looked at some of the challenges that have existed and still exist. And you can rest assured: there’s plenty of work left to be done in the future. This applies to both the journalistic and factchecking community, as well as the research sector.

Which is the perfect key word for what remains to be addressed: research.

If you are interested in a deeper dive into our research work, we very much invite you to our next session which takes place on 24 June 2025, again from 2-5 pm CET. You can still register for the webinar that focuses on hardcore research work by going to the section of the veraAI website that takes you to the registration page.

This said: on behalf of the entire vera.ai project consortium we thank you all very much for your attendance today! Stay happy and keep verifying!"

Recording 

Below, we provide a recording of the entire webinar. This is for those who missed out on the live event, and those who want to revisit some of the presentations and talks. Duration in total: just under three hours.

For your orientation we have also included the day's agenda. This should make it relatively easy to jump to individual sections should you wish to do so.

Happy viewing!

Author: Jochen Spangenberg (DW)