Authors

Eunice Esi Essuman, Nourhan Ameen

Abstract

LinguaShield addresses deceptive language used in fraudulent recruitment and otherexploitation-related offers. It is framed as a linguistic early-warning system that helps peopleexamine messages before paying fees, surrendering documents, travelling, or acceptingpurported work. The project focuses on prevention rather than post-harm reporting and isintended for students, job seekers, migrants, NGO case workers, labour inspectors, and othersevaluating opportunities. Its central premise is that coercion and trafficking risks can be mademore legible by identifying explainable linguistic warning patterns.

The prototype is organised as a connected suite rather than a single classifier. Its risk detectoraccepts pasted text and produces a risk verdict, confidence information, named signals,highlighted trigger phrases, a narrative forecast, a safer rewrite, and questions a recipient cansend back to verify an offer. A recruitment-journey visualiser was described as locating amessage within documented escalation pathways.

Complementary modules broaden the prototype from detection to learning and support. Fiveinteractive awareness stories present simulated choices and ask users to try again after anunsafe response. The evidence timeline accepts uploaded documents, extracts possibleexploitation signs into chronological order, and supports summary download or export.Community intelligence permits anonymous submissions or reading others’ reports and feedsa pattern explorer using phrase clusters and similarity scores. A chatbot answers predefinedquestions about flags, verification, and help. The safety area presents worker rights, reportingroutes, NGO contacts, and local resources; the demo selected Ghana from a five-countryinterface.

The modules were presented as mutually linked, allowing a user to move from analysis toevidence, education, or assistance. The team said the prototype could be downloaded as afile and opened on a phone without organisational infrastructure. Their proposed sustainabilitymodel keeps the public-facing detector, awareness stories, and community intelligence freefor opportunity seekers, while pursuing grants and future B2B API licensing for job boards,recruitment platforms, labour portals, and safe-migration programmes.

license

Creative Commons BY-SA 4.0