The Security Screening Automation Project is an artificial intelligence system being developed by the Canada Border Services Agency to help expedite the review of security screening referrals. The system uses a large language model to analyze referral data and create narrative summaries that highlight potential security concerns, assisting screening officers in their review process.
The system processes narrative text data from security referrals and identifies pre-determined areas of concern. Rather than making final decisions about security clearance, the AI generates summaries and flags potential issues for human screening officers to review and make final determinations. This approach is designed to help officers work more efficiently while maintaining human oversight of all security decisions.
This system is currently in development and will process personal information. It is being used by Government of Canada employees who conduct security screening reviews. The system relies on open-source large language models to generate its summaries and analytical output.
