Overview
- Experts at the TechSangam session in Lucknow on Friday urged caution about using AI in forensic work and said unverified AI outputs could jeopardise investigations and trials.
- Panelists said AI can speed analysis of fingerprints, DNA and CCTV by sifting large volumes of data to identify patterns and links that humans might miss.
- Speakers warned that deepfakes, manipulated digital files and biased or poor-quality training data can produce misleading results that would misdirect investigations or weaken courtroom evidence.
- The panel recommended specific safeguards including authentication of original evidence, full documentation of analytic procedures, explainable methods, independent validation of AI-assisted results and ongoing human oversight.
- Discussants called for capacity building across forensic, cyber and legal communities so investigators can detect AI errors, explain methods to courts and ensure results meet legal admissibility standards as the three-day International Dialogue continues.