Introduction

Forensic personal identification represents one of the core objectives of forensic science, particularly in cases involving unidentified human remains, mass disasters, and criminal investigations. Establishing the identity of deceased individuals is essential for legal investigations and humanitarian reasons, including disaster victim identification (DVI) and providing closure for families. The reconstruction of the biological profile—typically including sex, age, ancestry, and stature—plays a fundamental role in the identification process, especially when soft tissues are absent or when bodies are severely decomposed, fragmented, or skeletonized [1].

Traditionally, forensic identification relies on multiple complementary disciplines, including forensic anthropology, forensic odontology, and forensic genetics. Dental structures and craniofacial features are particularly valuable due to their durability and individual variability, making them useful for identifying victims in both criminal investigations and mass disaster scenarios [2]. In parallel, DNA profiling has become the gold standard for human identification, providing highly reliable genetic evidence that can link biological samples to individuals with high statistical confidence [3]. Additionally, morphological and metric analyses of skeletal structures remain fundamental tools in forensic anthropology for estimating biological characteristics when genetic material is unavailable or degraded [4].

In recent years, the rapid development of computational technologies has introduced new possibilities for forensic investigations through the application of artificial intelligence (AI). AI is a broad term describing computational systems capable of performing tasks that typically require human intelligence. Within AI, machine learning (ML) refers to algorithms that learn patterns from data to make predictions or classifications without explicit programming. Deep learning (DL) represents a subset of machine learning based on multi-layer artificial neural networks, particularly effective for analyzing complex imaging and high-dimensional datasets. [5]. These techniques have already demonstrated significant potential in several fields of medicine and science, enabling automated analysis of complex biological, imaging, and genetic data [6].

The integration of AI into forensic sciences is gaining increasing attention, with applications spanning forensic pathology, crime scene analysis, forensic radiology, and human identification [7]. Machine learning algorithms can assist forensic experts by identifying patterns in large datasets, improving classification accuracy, and supporting decision-making processes while potentially reducing subjective bias [8]. In particular, AI-based approaches have been explored in various identification-related domains, including facial reconstruction, skeletal analysis, genetic profiling, and microbiome-based identification [9, 10].

Despite the growing number of studies investigating the application of AI in forensic identification, the available evidence remains dispersed across different forensic disciplines and methodological approaches. Existing studies often focus on specific domains—such as forensic genetics, anthropology, or imaging—without providing a comprehensive synthesis of AI applications across the broader context of forensic personal identification. Therefore, a systematic evaluation of the current literature is needed to better understand the potential, limitations, and methodological challenges associated with the implementation of AI in this field.

Therefore, the aim of this systematic review was to analyze and synthesize the available evidence regarding the application of AI techniques in forensic personal identification.