Using AI in Real-Time Social Media Analytics: How OSINT Can Be Used to Detect Drug Trafficking Activities
Department of Artificial Intelligence and Data Science, Ajeenkya DY Patil School of Engineering, Pune, India · India
Abstract
More and more, drug trafficking groups have been communicating via public social media sites under layers of coded language, hashtags, emojis, and images. Monitoring such a huge variety of online content manually is a slow and laborious task, especially when phrases and patterns change. This research presents Artificial Intelligence (AI) and Open-Source Intelligence (OSINT) as means of helping the examination of published material from social media in an effort to detect content potentially needing further inspection. The methods covered by the study are the application of Natural Language Processing (NLP) for analysing written language and evaluating what a person is feeling or planning, the use of Computer Vision for the exploration of pictures that are publicly available, and machine-learning for the labelling of a suspicious content or a piece of material. A conceptual AI–OSINT workflow diagram is illustrated, starting with data acquisition, going through preprocessing, then multimodal analysis, followed by risk scoring, ending with human reviewing. According to the research papers and seminar content, the challenge of fusing both image and text analyses, responding to the development of secret language, having large enough databases, and producing clear explanations are still major issues. As the internet material can sometimes be incomplete and an automatic alarm system cannot be used to confirm a crime has been committed, AI-supported OSINT should be seen as a decision aid that is backed by privacy safeguards, a fair assessment, and the human double-check.
Keywords
How to Cite This Article
Rights
BODHIVRUKSHA JOURNAL OF DIVERSE DISCIPLINE