Revolutionary AI framework reshapes forensic age estimation using dental records

A new AI framework could transform how forensic experts estimate age using dental records. Developed by Palmela Pereira and presented in 2026, the system aims to address long-standing challenges in accuracy and reliability. Traditional methods often struggle with bias and inconsistency, but this technology promises a more objective approach.

The study was discussed at an event hosted by the PTS – Institut für Fasern & Papier on November 11, 2025, with contributions from specialists like Dr. Georg Dietz on advanced AI techniques.

Forensic age assessment has always been a difficult task, particularly when dealing with living individuals. Errors can have serious legal consequences, especially in cases involving asylum-seeking minors. Pereira's research introduces a system that uses convolutional neural networks (CNNs) to analyse dental X-rays and match developmental stages with probabilistic age estimates.

The AI does not just provide an age prediction—it also includes confidence scores. This allows human experts to assess reliability before making final decisions. By automating initial evaluations, the technology speeds up forensic workflows and reduces the workload on specialists.

One key advantage is the system's adaptability. Built with transfer learning, the AI can adjust to different populations, improving its usefulness worldwide. However, the study stresses that ongoing validation and refinement are essential to maintain accuracy and fairness.

Ethical considerations also play a major role. The research highlights how AI can bring transparency to age assessments, ensuring results are both replicable and accountable. This is particularly important in humanitarian contexts, where incorrect age determinations can impact vulnerable individuals.

The new AI-driven method offers a way to improve forensic age estimation, combining speed with greater precision. Its ability to provide confidence metrics and adapt to diverse groups could make it a valuable tool in legal and humanitarian settings. Further testing and development will determine how widely it can be applied in real-world cases.