Reality, Rewritten: AI as Forger and Investigator#

Dominik Haenni - Zurich Forensic Science Institute

• Why: The opening keynote frames the question the three agentic-AI talks later dance around: when the same technology fabricates evidence and analyzes it, what happens to expert judgment and trust in organizations — which is, in the end, the same question as trusting an agent's root-cause analysis on a production network.
• What: An exploration of generative AI's dual role — producing convincing fake images, cloned voices, and plausible documents, while also accelerating forensic analysis, research, and software development — with current examples from the Institute's casework, and a look beyond the tool at responsibility and how organizations work when AI compresses timelines.
• How: Case-example driven (the abstract does not describe a method); the argument is that when AI "changes the rules," business as usual in evidence handling — and by extension, any function that certifies truth — is not affordable.
• Where applicable: Not a networking talk, but its trust-in-automated-findings framing transfers directly to anyone shipping AI-generated analysis into operational workflows. (from abstract only)