DFRLab launched FIAT

The Atlantic Council’s Digital Research Laboratory launched the Foreign Interference Attribution Tracker or FIAT, to gain a better overview of USA election interventions.

The database and the analysis application based on it provide a visual overview of interventions that take place over time, outline their scope and impact, and provide access to the details of the case.

The database is based on identified cases that have been previously analyzed and described.

All cases entered must be related to external intervention (presumably in the digital environment), recent and current, and related to the 2020 USA elections.

The tool is based on two meterings: attribution impact and attribution rating.

The first measures the distribution of articles and content related to the case over the last seven days. Summarizes Facebook sharings, likes, and comments; sharings on Twitter and mentions on Reddit. The impact assessment of the case is based on the analysis tools; SerpApi, BuzzSumo and CrowdTangle.

However, the attribution score is compiled on the basis of 18 criteria that assess the messages to be disseminated.18 conditions are in turn divided into 4 major groups: reliability, objectivity, evidence and transparency.

Credibility takes into account, for example, whether the news source has a financial interest in the results of the message being disseminated and whether the message is unrelated to any political campaign. It also looks to see if the source is previously unrelated to any misinformation campaigns and many other similar terms.

The category of objectivity takes into account whether the message avoids one-sided and loaded wording or whether the title conveys the content of the message, etc.

The evidence assesses whether it is possible to identify information delivery methods, tactics and platforms, from the message as well as the individuals and countries behind it and other similar markers.

In terms of transparency, the focus is on the quality of the data, including whether the case can be identified from public sources, whether the method is described in sufficient detail, whether there is public access to the case information, etc.

For each attribution, a score from zero to 15 is prepared, with a higher number indicating the reliability of the entry.

The tool will initially be used for the US 2020 elections, but could be used during other countries’ election periods to gain a clear picture of attempts to influence them.

Photo: Screenshot of the tool’s website.