Democracy Reporting International (DRI) recently launched a user-friendly concept for a web-based toolkit to monitor social media during elections. As the phenomenon of social media influencing the political arena is something known only for about 5 years, little knowledge existed about the impact on democracy. With their project, ‘Testing (EU Designed) Methodology of Media Monitoring in Elections’ that was funded by Civitates and done in cooperation with MEMO 98, leading and which led to the development of the toolkit, DRI seeks to significantly improve the real-time understanding of social media dynamics around elections.

Who in particular has DRI developed this toolkit for?

The toolkit is focused on being used by civil society organisations (CSOs) primarily. Through the toolkit, their understanding of social media’s impact on democracy will increase. Next to that, they will be able to learn a lot of things about monitoring elections, for instance about what to do with the data collected. The observation of elections can, however, be done by everyone who has an interest: researchers, watchdog organisations, or an election observer from an ‘official’ organisation. While the toolkit does make research easier, it fits better with established organisations, as it demands quite some effort and skills.

Have you received feedback from CSOs who worked with the toolkit already?

We received feedback from election observers from the EU who expressed to have found it very useful. Monitoring social media is a new field which wasn’t part of the usually assigned missions that the election observers usually carry out. DRI is part of a consortium with GIZ and MEMO98 which won the EU’s bid to work with the Electoral Observation and Democracy Support (EODS) to train election observers from the EU. It is quite some change to go from regular media monitoring about elections to social media monitoring. The toolkit that we provided is a kind of recipe book that gives directions on how to approach this monitoring from various angles. The official EU observers, however, need to work with one single standarised methodology, something we are also discussing now.

DRI worked with five different civil society groups in Europe on this project. How was that work related to the toolkit?

Alongside the development of the toolkit, we supported five civil society groups to monitor social media during elections in their respective countries. This included training, general support, and providing methodological assistance, feedback, and report review. When we started working with them, we only had the guide that we had written ourselves about monitoring social media during elections, which was actually the first of its kind. Working with the five CSOs made us understand that there are different ways to approach the issue. Some of our partners only looked at what official actors would post, others monitored online political advertisement. We incorporated all of these lessons in the toolkit, as well as in a lessons learned paper. We conducted a survey at the end of the project that gave us valuable insights and showed us what the pros and cons of our approach were. The variety of approaches really enriched the content of the toolkit.

Was there anything remarkable that the CSOs discovered or achieved with their monitoring?

The Croatian partners were reporting on political advertisement. They did good research, with which they aimed to understand whether politicians accurately reported political advertisement expenditures on Facebook in their financial reports. During their analysis, the team found an error and directly reported this. The concerned political candidate corrected the error in their reporting. It is difficult to claim that this was a direct result from the reporting of our Croatian partner, but it is obvious that they found something that hadn’t been noticed before. The Portuguese partners we worked with, looked at the role of groups on Facebook spreading false information before the elections. They found that narratives from a few years before were being recycled to misguide users. By shedding light on that matter, they created awareness among the people about how they are manipulated.

What do you personally see as the biggest achievement of this project?

DRI started creating this methodology to actually understand what was happening on social media, and providing a space where this information is easily accessible is a big step towards more transparency. As we didn’t have all the data necessary for monitoring, as tech companies are still lacking the provision of all the necessary transparency, we could not properly scrutinise elections. With this tool, we can now point out exactly how to collect data for research, since each platform has its own data access rules and procedures. We also point out the level of difficulty and technical knowledge needed to perform this activity, as well as indicating what data is missing to properly scrutinise elections. As we were performing the work during the EU’s discussion on regulatory approaches to tech companies, the learnings of the Civitates funded project had an indirect impact on our contributions to the EU regulatory process.

What was the biggest challenge for your project with regard to COVID-19?

When the crisis hit Europe, we were already in the last phase of our project. We only had to shift one planned event into an online meeting. Furthermore, not so much changed for us. The toolkit, being a digital guide, was actually launched in time. It is an interactive and easy-to-use tool, which facilitated further online training we conducted in the second half of 2020.

You also recently published a report about deep fakes. Can you tell us more about that?

The paper we wrote gives information about what has been developed so far and what the current trends are. Artificial Intelligence creating videos is a new development and example of how technology can be used for the better, but also be abused in future disinformation campaigns. The film industry can really benefit, for instance, as you can dub an actor’s lips very naturally. There are many good usages, but as with almost everything, it can also be misused as well. If a politician is mimicked, this may have a big impact on democracy as people might start to disbelieve what they see. It can also be used as an excuse, for instance, if politicians claim that what they said, was not really said by them, but the result of a deep fake. As the toolkit, being an online product, can easily be updated, we will add methodologies of how to approach deep fakes and other new technology to the toolkit in the future as well.