FALC: An Artifact-Centric Life Cycle Model and Research Agenda for AI Forensics
Organizations: Chair for Cybersecurity, University of Augsburg, 86159 Augsburg, Germany · Chair for Security and Privacy in Socio-technical Systems, Ruhr-University Bochum, 44801 Bochum, Germany
Abstract
AI systems increasingly appear in criminal investigations as tools, targets, and sources of evidence, yet AI forensics lacks a systematic account of the artifacts they produce: investigators often do not know which traces to look for, and for many traces they do find, no established analysis technique exists. In this paper we first developed the Forensic AI Life Cycle (FALC) that unlike other life cycle models shifts the focus onto the forensic artifacts created during the life cycle of an AI system. Through FALC we derived a reference list of artifacts that might serve an investigation. Second, we collected examples for crimes with AI involvement throughout FALC and questions likely to be asked when investigating these crimes. Third, we propose a taxonomy of forensic fields of expertise with a focus on the current possibility to analyze AI forensic artifacts using established techniques. For investigators, the model provides a systematic basis for identifying which artifacts to look for when an AI system is relevant to a case. For researchers, the classification shows where existing forensic methods already suffice and where new techniques are most urgently needed.
Figures & tables
| Stage | Forensic Question | Relevant Artifacts | Obtain | Analysis |
|---|---|---|---|---|
| Governance and Quality Assurance | Determining accountability | policies and procedures, licenses, design documents, audit results, test procedures, test results, monitoring | est | est |
| Governance and Quality Assurance | First occurrence of abnormal behavior | test results, audit results, user feedback, backups, and version history | est | est |
| Data Acquisition | Data source and acquisition method | networking, access logs, inspection, generation system, negotiations and applications, monetary transactions | est | est |
| Data Acquisition | State of the data when acquired | raw data, inspection, storage, access logs, generation system | est | adapt |
| Data Preparation | Data modifications | prepared data, partially prepared data, processing pipeline, collaboration artifacts, operating system artifacts | est | adapt |
| Data Preparation | Data modifications | AI system, augmentation input | adapt | new |
Appendix figures & tables2 assets
Supplementary material from the paper’s appendix.
Appendix
| Governance and Quality Assurance | Data Acquisition | Data Preparation | Model Development | Model Integration | Deployment | Use | ||
| AI Training Forensics | Training Process Forensics | X | ||||||
| Dataset Forensics | X | X | ||||||
| Environment Forensics | X | X | ||||||
| AI Substrate Forensics | X | X | X | X | ||||
| AI Application Forensics | X | X | ||||||
| AI Model Forensics | Model Authentication Forensics | X |
| Docket No. | AI related artifact | FALC stage | FALC artifact group |
|---|---|---|---|
| 4 | Sensors used by automation system | Governance and Quality Assurance | documentation |
| Model Integration | before the model | ||
| 4 | Sensor calibration schedule and last calibration | Governance and Quality Assurance | policies and procedures |
| 4 | Procedure to plan a route | Governance and Quality Assurance | policies and procedures |
| 4 | ADS engagement procedure and restrictions | Governance and Quality Assurance | policies and procedures |
| 4 | ADS disengagement procedures and triggers | Governance and Quality Assurance | policies and procedures |
Digital Forensics
Digital forensics develops methods to recover, authenticate, and interpret digital evidence while preserving its integrity for investigations.
See all 75 papers