Mobasi at DFRWS 2026 – Perspectives on Applying AI in Digital Forensics
DFRWS published the report from a practitioner-led session on AI in digital forensics casework, with Eoghan Casey, Jim Jones, Jessica Hyde, Cameron Kelley, and Chad Steel.
Mobasi
At the Digital Forensics Conference 2026 Americas, organized by DFRWS at George Mason University, Mobasi's Cameron Kelley joined a practitioner-led working session on using AI in digital forensics casework. DFRWS has now published the report from that session.
Eoghan Casey, Jim Jones, Jessica Hyde, Cameron Kelley, and Chad Steel each shared their experience and led group discussions, guided by live audience polls that kept the focus on what practitioners in the room were seeing.
What the session covered
- How to evaluate the performance of AI tools
- Appropriate and inappropriate uses of AI in digital forensics
- Grounding AI in digital forensics expertise to improve reliability and accuracy
- AI engineering and guardrails
- Investigators and attorneys analyzing digital evidence with AI on their own
Three themes
Practitioners own the result. Digital forensics practitioners are responsible for verifying what AI-assisted and AI-generated tools produce. No one should push a button and accept the output, or hand it to a non-specialist to judge.
Treat AI like any other forensic tool. That means being able to audit what was done and trace every result back to its source data.
Bring investigators and attorneys in. Guidance written only for practitioners won't reach the investigators and attorneys who now use these tools, which leaves them open to errors and wrong conclusions. Practitioners need to work with them on responsible use and stay part of interpreting and validating the evidence.
These are the standards we build Mobasi to: every finding links to the record behind it, so an examiner can check it and stand behind it.
Read the full report: Practitioner Perspectives on Making AI Defensible in Digital Forensics Casework (PDF, DFRWS).