The Government Accountability Office (GAO) has presented three possible actions policymakers can take to address challenges related to the use of forensic algorithms by law enforcement agencies and one is supporting increased training for criminal investigators and analysts.
GAO noted that training on human factors could help reduce risks related to analyst error and enable users to understand and interpret the results, according to a report published Tuesday.
With regard to facial recognition and latent print algorithms, providing analysts training on cognitive biases could help improve objectivity and awareness.
GAO also recommended that policymakers support the development and implementation of policies and standards on the appropriate use of such algorithms and advance increased transparency with regard to the use, performance and testing of the technology.
Law enforcement agencies use latent print, facial recognition and probabilistic genotyping algorithms in criminal investigations. However, the congressional watchdog found that the use of forensic technologies presents challenges to investigators and analysts, including potential bias and misuse and difficulty in interpreting results.