情報処理学会 第86回全国大会 会期:2024年3月15日~17日

4U-01
刑事司法におけるディープフェイク検出システムの予備的なユーザ評価について
○Hydara Ebrima,菊地真人,大囿忠親(名工大)
As the prevalence of deepfake technology continues to rise, its potential impact on criminal justice raises significant concerns. This paper presents a preliminary user evaluation of a deepfake detection system for facial evidence verification in criminal justice. The evaluation unveiled the efficacy of the system in identifying deepfake facial evidence through prediction performance and user perception confidence. Through user feedback and performance metrics, the paper sheds light on the system’s strengths and limitations in a real-world setting. The findings contribute valuable insights to the ongoing research on the use of artificial intelligence for facial evidence verification in criminal justice.