Can AI-generated faces serve as fillers in eyewitness lineups?

Area of Science:

• Forensic Psychology
• Artificial Intelligence in Law Enforcement

Background:

• Eyewitness identification is crucial in criminal investigations.
• Photo lineup construction, particularly selecting fillers, is challenging for law enforcement.
• Fillers should resemble the suspect to avoid biasing the identification.

Purpose of the Study:

• To investigate the feasibility of using artificial intelligence (AI) to generate filler images for eyewitness photo lineups.
• To assess whether participants can differentiate between real suspect photos and AI-generated filler images.
• To evaluate the impact of AI-generated fillers on eyewitness identification accuracy.

Main Methods:

• AI algorithms were used to generate images of individuals to serve as fillers.
• Participants were presented with photo lineups containing a real suspect and AI-generated fillers.
• Identification accuracy and the ability to detect AI-generated images were analyzed.

Main Results:

• Participants largely failed to detect the differences between real suspect photos and AI-generated fillers.
• No significant differences were observed in identification outcomes between lineups with real and AI-generated fillers.
• The study focused on white, male targets, indicating a need for further research on generalizability.

Conclusions:

• AI-generated images show promise as a feasible tool for creating effective filler images in photo lineups.
• The use of AI in lineup construction could potentially improve efficiency and reduce bias.
• Further research is recommended to validate these findings across diverse demographics and contexts.