Excellence in Research: Enhancing the Trustworthiness of LLM and VR Assisted Digital Forensic Investigation

This project enables the trustworthy, interpretable, and effective integration of LLMs into digital forensic workflows. It investigates the sources of hallucinations in forensic contexts, develops techniques to minimize fabricated or inaccurate outputs while ensuring reliability aligned with evidentiary standards, enhances the interpretability of LLM-generated outputs through explainable AI techniques, and design and validate performance metrics and visualization-based representations that provide systematic evaluation of LLM reliability in digital forensic applications. The PIs also develop a complementary virtual reality (VR) training environment that enables users to interact with forensic scenarios and evaluate LLM-generated evidence in context. This research advances foundational knowledge at the intersection of artificial intelligence, digital forensics, and legal reasoning by addressing the trustworthiness of LLMs in high-stakes investigative contexts. The project combines hallucination mitigation, interpretability techniques, and formal evaluation frameworks with immersive visualization and user-centered design, the project contributes to the development of evidence-grade, transparent AI systems. The interdisciplinary approach—spanning AI, forensic science, legal processes, and human-computer interaction—provides innovations that strengthen forensic investigations and extend to other critical domains such as cybersecurity, healthcare, and law enforcement. This project is funded by the U.S. National Science Foundation (NSF) Historically Black Colleges and Universities-Excellence in Research (HBCU-EIR) Program. HBCU-EIR supports research at public and private HBCUs to strengthen their research capacity and promote engagement with NSF. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.