A Preliminary Study of AI-Driven Scams in Algeria: Exposure, Detection, and a Roadmap

The rapid spread of generative AI has created a new class of digital fraud. Deepfakes, voice clones, and AI-written content are now widely available and difficult to distinguish from real media. This study examines AI-driven scams in Algeria using a bilingual online survey of +292 participants, a review of current generative tools, and targeted case studies. The survey shows high exposure to AI-generated promotional or impersonation content, while many respondents struggle to identify manipulated media. Only a small share reported formal cybersecurity training, and a nontrivial portion admitted to interacting with suspicious links. The paper documents common attack patterns, lists representative tools, and proposes a staged detection roadmap combining provenance, fast heuristics, and deeper forensic analysis. The article includes anonymized user accounts and fabricated image case studies used for demonstration and awareness. It concludes with concrete recommendations for short term awareness campaigns, medium term dataset and tool development, and long term legal and research investments. This work is a preliminary, practice-oriented contribution intended to inform policy makers, educators, platform operators, and researchers working on Arabic and North African contexts.