The effects of lineup size on the processes underlying eyewitness decisions

Self Cite

Recent advances in artificial intelligence (AI) enable the generation of realistic facial images that can be used in police lineups. The use of AI image generation offers pragmatic advantages in that it allows practitioners to generate filler images directly from the description of the culprit using text-to-image generation, avoids the violation of identity rights of natural persons who are not suspects and eliminates the constraints of being bound to a database with a limited set of photographs. However, the risk exists that using AI-generated filler images provokes more biased selection of the suspect if eyewitnesses are able to distinguish AI-generated filler images from the photograph of the suspect’s face. Using a model-based analysis, we compared biased suspect selection directly between lineups with AI-generated filler images and lineups with database-derived filler photographs. The results show that the lineups with AI-generated filler images were perfectly fair and, in fact, led to less biased suspect selection than the lineups with database-derived filler photographs used in previous experiments. These results are encouraging with regard to the potential of AI image generation for constructing fair lineups which should inspire more systematic research on the feasibility of adopting AI technology in forensic settings.

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

“…As in our previous experiments 13 , 14 , 28 , 29 , there were two mock-crime videos that were parallel to each other. In each of the two videos, four different culprits were shown.…”

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

See 2 more Smart Citations

On the advantages of using AI-generated images of filler faces for creating fair lineups

Bell,

Menne,

Mayer

et al. 2024

Self Cite

“…When the 2-HT eyewitness identification model was used to evaluate the effects of lineup size on the latent processes underlying eyewitnesses’ responses to lineups 19 , interesting findings emerged on which we build here. Reducing the size of lineups (from six to three or from five to two persons) resulted in two remarkable outcomes: The probability of culprit-presence detection (represented by model parameter dP ) was significantly higher and the probability of guessing-based selection (represented by model parameter g ) was significantly lower in smaller compared to larger lineups.…”

Section: Introductionmentioning

confidence: 95%

On the possible advantages of combining small lineups with instructions that discourage guessing-based selection

Therre,

Bell,

Menne

et al. 2024

Self Cite

The primary argument for including large numbers of known-to-be innocent fillers in lineups is that guessing-based selections are dispersed among a large number of lineup members, leading to low innocent-suspect identification rates. However, a recent study using the two-high threshold eyewitness identification model has demonstrated advantages of smaller lineups at the level of the processes underlying the observable responses. Participants were more likely to detect the presence of the culprit and less likely to select lineup members based on guessing in smaller than in larger lineups. Nonetheless, at the level of observable responses, the rate of innocent-suspect identifications was higher in smaller compared to larger lineups due to the decreased dispersion of guessing-based selections among the lineup members. To address this issue, we combined smaller lineups with lineup instructions insinuating that the culprit was unlikely to be in the lineup. The goal was to achieve a particularly low rate of guessing-based selections. These lineups were compared to larger lineups with neutral instructions. In two experiments, culprit-presence detection occurred with a higher probability in smaller compared to larger lineups. Furthermore, instructions insinuating that the culprit was unlikely to be in the lineup reduced guessing-based selection compared to neutral instructions. At the level of observable responses, the innocent-suspect identification rate did not differ between smaller lineups with low-culprit-probability instructions and larger lineups with neutral instructions. The rate of culprit identifications was higher in smaller lineups with low-culprit-probability instructions than in larger lineups with neutral instructions.

Lineup position affects guessing-based selection but not culprit-presence detection in simultaneous and sequential lineups

Mayer,

Bell,

Menne

et al. 2024

The two-high threshold eyewitness identification model was applied to investigate the effects of lineup position on the latent cognitive processes underlying eyewitness responses in lineups. In two experiments with large sample sizes and random assignment of culprits and innocent suspects to all possible lineup positions, we examined how detection-based and non-detection-based processes vary across the positions of six-person photo lineups. Experiment 1 (N = 2586) served to investigate position effects in simultaneous lineups in which all photos were presented in a single row. Experiment 2 (N = 2581) was focused on sequential lineups. In both experiments, lineup position had no effect on the detection of the presence of the culprit. Guessing-based selection, in contrast, differed as a function of lineup position. Specifically, a lineup member placed in the first position in a lineup was significantly more likely to be selected based on guessing than lineup members placed in other positions. These results justify the practice of avoiding to place the suspect in the first position in a lineup, as this placement increases the suspect’s chance of being selected based on guessing.