This study explores the potential of AI to support the work of military intelligence analysts.In the study, 30 participants were randomly assigned to an experimental condition (in which they could use a proprietary AI tool) and a control group (no access to the AI tool). In both conditions, participants had access to the same dataset of 50 media articles and were asked to provide a comprehensive picture in response to a series of realistic military intelligence tasks. The proprietary AI tool included text search, automatic text summarization, and named entity recognition (NER) capabilities. It was shown that under time pressure, the use of the AI features resulted in better assessments than the control group. It was also shown that the probability estimates of the experimental group were closer to those of the experts. Despite these demonstrably better analysis results and probability estimates in the experimental group, no higher confidence in the sources used for the analysis task was found. Finally, the paper identifies the limitations of using AI in military intelligence, particularly in the context of analyzing ambiguous and contradictory information.      «

This study explores the potential of AI to support the work of military intelligence analysts.In the study, 30 participants were randomly assigned to an experimental condition (in which they could use a proprietary AI tool) and a control group (no access to the AI tool). In both conditions, participants had access to the same dataset of 50 media articles and were asked to provide a comprehensive picture in response to a series of realistic military intelligence tasks. The proprietary AI tool inc...     »