Conspiracy Frame: a Semiotically-Driven Approach for Conspiracy Theories Detection
This work addresses the challenge of understanding and detecting conspiracy theories, which can cause social conflict, but it is incremental as it builds on existing frame-semantics and semiotics approaches without achieving clear performance gains.
The authors tackled the problem of detecting conspiracy theories by introducing the Conspiracy Frame, a semantic representation derived from frame-semantics and semiotics, and the Con.Fra. dataset of annotated Telegram messages. They found that while injecting frames into LLMs did not clearly boost performance, it showed potential and revealed abstract semantic patterns like 'Kinship' and 'Ingest_substance' for more aware detection.
Conspiracy theories are anti-authoritarian narratives that lead to social conflict, impacting how people perceive political information. To help in understanding this issue, we introduce the Conspiracy Frame: a fine-grained semantic representation of conspiratorial narratives derived from frame-semantics and semiotics, which spawned the Conspiracy Frames (Con.Fra.) dataset: a corpus of Telegram messages annotated at span-level. The Conspiracy Frame and Con.Fra. dataset contribute to the implementation of a more generalizable understanding and recognition of conspiracy theories. We observe the ability of LLMs to recognize this phenomenon in-domain and out-of-domain, investigating the role that frames may have in supporting this task. Results show that, while the injection of frames in an in-context approach does not lead to clear increase of performance, it has potential; the mapping of annotated spans with FrameNet shows abstract semantic patterns (e.g., `Kinship', `Ingest\_substance') that potentially pave the way for a more semantically- and semiotically-aware detection of conspiratorial narratives.