top of page

AI Forensics Workshops: Practical Detection, Verification, and Reporting

Hands‑on training that teaches your team how to spot, test, and document AI‑generated or manipulated media using real cases, multi‑model detectors, classic forensics, and clear reporting that stands up to scrutiny.

WHO THE WORKSHOP IS FOR?


Editors and fact-checkers; investigative and OSINT teams; law enforcement and cybercrime units; risk, fraud and compliance in finance; brand protection and trust and safety teams; legal, PR and crisis communications.

 

LEARNING OUTCOMES

​

  • Understand the threat landscape and why single detectors fail.

  • Learn how modern generators work and the artifacts they leave.

  • Follow a proven workflow that blends multi-model AI detection with classic image forensics and contextual checks.

  • Run and interpret: ensemble detectors, EXIF and XMP review, ELA, noise and CFA analysis, resampling, JPEG ghosts, lighting checks and face-region scrutiny.

  • Build reverse-image timelines, hunt variants and perform cross-angle comparisons.

  • Apply geolocation, weather and physical-plausibility checks for high-stakes claims.

  • Triage video and deepfakes: frame sampling, lip sync, blink cadence, motion consistency, compression tells.

  • Understand watermarks and provenance: model watermarks, invisible marks, C2PA basics, current strengths and gaps.

  • Maintain chain of custody, hashing and evidence handling.

  • Write clear verdicts with confidence scoring, residual risk and next steps.

 

FORMATS

​

  • Executive briefing – 90 minutes: strategic overview, live demos, decision checklist.

  • Half-day – 3.5 hours: core workflow plus two guided labs.

  • Full day – 7 hours: complete workflow, four labs and a graded capstone case.

  • Two-day certification – 12 to 14 hours: deep labs, red-team vs blue-team exercise and a final report graded against a rubric.

 

TYPICAL FULL-DAY AGENDA

​

  1. Orientation and threat overview

  2. Detectors: strengths, failure modes, ensemble scoring

  3. Classic image forensics: artifacts and manipulation tests

  4. Reverse search and timeline building

  5. Contextual checks: geo, weather, uniforms, reflections, depth

  6. Video and deepfake triage

  7. Chain of custody and reporting

  8. Capstone lab with live debrief

 

HANDS-ON LABS

​

  • Lab 1: AI vs camera capture on compressed social files

  • Lab 2: Splice and content-aware fill detection on DSLR images

  • Lab 3: Reverse-search timeline with variant clustering

  • Lab 4: Deepfake clip triage and frame evidence pack

  • Capstone: mixed set with conflicting signals, full report and presentation

 

TAKE-HOME TOOLKIT


Workflow poster; image and video triage checklists; confidence scale and wording guide; report template with evidence tables and hash blocks; escalation matrix for publish, takedown or referral; vendor and model evaluation checklist.

 

DELIVERY OPTIONS


Live online via secure video; on-site at your office or training center; air-gapped or on-prem for sensitive units. NDA available. Custom datasets and private case workups on request.

 

REQUIREMENTS


Laptop with admin rights to run provided tools; stable internet for online sessions. Optional GPU machine if you want to run local models, or we can supply hosted access.

 

CERTIFICATION AND CREDITS


All participants receive a CertiSight AI certificate of completion. Two-day courses include an assessment report suitable for CPD credit applications with professional bodies.

 

DATA PROTECTION


We avoid uploading personal data to third parties where possible. Training evidence is de-identified or publicly sourced. Client-supplied cases are handled under NDA with chain of custody and deletion on request.

 

PRICING


Pricing depends on format, group size, sensitivity, travel and any custom datasets or on-prem setups. Volume and multi-team discounts are available.

​

bottom of page