BetXplain: explanation-annotated dataset for detecting manipulative betting ads on social media

arXiv·low signal

This arXiv paper (2026-06-25) introduces BetXplain, a new explanation-annotated dataset for detecting manipulative and deceptive betting advertisements on social media, addressing the lack of public labeled data in a space where persuasive tactics can encourage risky behavior. The annotations target both detection and human-readable explanations of the manipulation. It's a useful resource for trust-and-safety and content-moderation builders working on persuasive-harm detection.