The DRAGON-Spotter tool integrates Linguistics and Artificial Intelligence (AI) to detect online grooming content, pinpointing the manipulative language tactics that groomers use, from making children feel emotionally isolated to communicating sexual intent implicitly and explicitly to them. DRAGON-Spotter is designed to aid law enforcement’s online grooming detection work and can therefore have a significant impact on our ability to keep children safe from online grooming.
The tool can be installed provides a platform for law enforcement to upload obtained conversations to be processed by the DRAGON-Spotter machine learning model. The tool uses this model to produce several scores, which are used for the classification of a whole conversation as online grooming/non-grooming and detects occurrences of groomer tactics within a conversation. Output results are analysed by the user and used to streamline detection work, such as:
- Triaging and prioritising cases to process;
- Reducing analyst/officer workload in processing cases;
- Strengthening an existing case;
- Identifying new cases.
Users of DRAGON-Spotter can interpret the tool’s results and integrate them into their working practices in an informed way. Our informative user guide is available to aid navigating the tool, explaining results and ensuring ease of use.
Following a period of rigorous development and testing, working with law enforcement departments and safeguarding practitioners internationally, Project DRAGON-S is conducting a thorough evaluation of both the DRAGON-Shield and DRAGON-Spotter tools in 2023.
If you are a Law Enforcement Agency interested in hearing more about our DRAGON-Spotter tool and would like to contact us, please email project.dragons@swansea.ac.uk.