Introduction
The GetResilience app is to enable mass-flagging of illegal and potential harmful content. The app needs to operate under two major constraints:
- The DSA Regulation
- The objectives and planned outcomes of project Saufex
Saufex objectives
- Standardise knowledge, building on existing good practices to further develop a common and standardised framework for FIMI analysis;
- De-centralise and democratise processes around FIMI analysis and response;
- Incorporate community-driven quality assurance processes and expand theoretical understandings of FIMI within the defender community.
Saufex outcomes
The objectives aim to achieve the following outcomes:
- Enable proactive detection, analysis, and countermoves in case of foreign attempts to undermine democratic processes and develop a better understanding of the Tactics, Techniques, and Procedures (TTPs) involved in these attempts.
- Create an improved response system for policymakers and other stakeholders so that they are better equipped at dealing with FIMI threats.
- Improve coordination among EU member states and international partners in combating FIMI and provide a better theoretical understanding of the tension between the tenets of democracy and FIMI.
Major dilemmas
The challenge for the app is to resolve the following major dilemmas:
- The app should have low entrance friction for end-users who wish to flag but at the same time result in useful flags, that is (1a) flagging illegal content or (1b) content or procedures that are considered by the flagger to be harmful content;
- The app should have low entrance friction for end-users who wish to flag but the app could face penalties for abusive reporting;
- The app should have low entrance friction for end-users but should not be a target for adversarial gaming, e.g. “mass-flagging” by bots or coordinated actors to censor legitimate speech;
- The DSA process requires a good faith statement but the app needs to allow for anonymous flagging;
- The DSA process requires a good faith statement which makes AI automation of the flagging flow that would enable mass access to flagging a challenge;
- The DSA requires flagger accountability and trustworthiness but that triggers GDPR risks;
- Non-anonymous flagging potentially triggers the GDPR but the app should be as automized as possible;
- Regular end-users are not trained in legal details but insufficient knowledge of the details potentially disqualifies the flag: e.g. reporting non-illegal FIMI as “illegal content” leads to “liability leakage” and regulatory rejection;
- The DSC is to be informed at an earlier stage and more comprehensively but the DSA flags are platform-oriented;
- The DSA is platform-oriented but FIMI and other threats are often cross-platform campaigns.