Background: Cervical cancer remains a leading cause of death among women and is largely preventable; the burden is most significant in low- and middle-income countries despite the availability of effective HPV vaccination and screening. Artificial intelligence (AI) could strengthen screening programs by improving accuracy, throughput, and access across cytology, colposcopy, and mobile app–based platforms.

Methods: Following PRISMA-ScR, we searched PubMed, EMBASE, MEDLINE, and CINAHL (2014–2025) for peer-reviewed studies applying Artificial Intelligence (AI)/ Machine Learning (ML)/ Deep Learning (DL) to cervical cancer screening or diagnosis using Pap/liquid-based cytology (LBC), colposcopy, or mobile tools, and reporting diagnostic or economic outcomes. Two reviewers independently screened, extracted, and adjudicated data.

Results: Of 724 records, 19 studies met inclusion criteria. AI-assisted Pap/LBC cytology for CIN2+ showed sensitivity ~86–90% and specificity ~51–95%; for CIN3+, sensitivity ~87–92% with specificity ~51–61%, often matching or exceeding conventional workflows while increasing slide throughput. Colposcopy AI achieved sensitivity/specificity trade-offs depending on thresholds (e.g., high-grade-or-worse: ~72%/94%; more sensitive setting: ~91%/52%), supporting targeted biopsy decisions and reducing operator variability. Mobile/point-of-care tools reported strong performance in validation settings (e.g., smartphone VIA ~99% sensitivity/~97% specificity; AI-supported HPV assays ~100%/~100%), with minimal hardware requirements and potential for taskshifting. Across modalities, AI commonly reduced time-to-diagnosis and expert workload. Eight studies assessed economics; AI-LBC at 5-year intervals was cost-effective (example ICER ≈ US$8,790/Quality-Adjusted Life Year(QALY)), and decision-support systems reduced unnecessary colposcopies, suggesting favorable value in resource-constrained settings.

Conclusions: AI-enabled screening can deliver clinically meaningful accuracy with operational gains that are likely to improve coverage and equity, particularly where cytology expertise and colposcopy capacity are limited. The greatest value appears where moderate-to-high diagnostic performance coincides with low deployment costs (e.g., AI-assisted cytology and mobile platforms). Evidence is promising but heterogeneous; real-world, prospective studies with standardized endpoints and full economic evaluations are still needed to confirm effectiveness, scalability, and affordability within national screening programs aligned to World Health Organization (WHO) elimination targets.