2025_EML Capstone Projects

Contingency plan: If false positives continuously keep disrupting, the plan is to temporarily revert to a stricter rule-based machine learning model and incrementally readjust machine learning parameters. Compliance and privacy risks: As the Veterans of Affairs handles sensitive personal information, using real data would significantly violate privacy which is secured under FISMA and HIPAA policies. Mitigation strategy: Fully synthetic data utilized to avoid potential privacy violation. Synthetic data reflects as real patterns but does not include any real user information. Also, audit trails have been created to capture inputs, decisions and model logic. A governance framework has been developed with compliance checkpoints which are also simulated. Contingency plan: Privacy impact assessment would be conducted before actual implementation with current VA SIEM which is a system to limit sensitive personal information stream. Adoption risks and trust: Stakeholders might be concerned due to the nature of the machine learning the black-box concept. The changes of resistance are high from leadership and compliance personnel as the system has complexities and is not very easy to interpret. Mitigation strategy: Adopting phase mentality mitigated the risks when started with established rules and followed by adding machine learning layers. Dashboard visualization and logs have been implemented to explain the future gravity of the system. Additionally, demo sessions are hosted for internal stakeholders to pursue detection outputs and logic. Contingency plan: If resistance from leadership and stakeholders remains, only the rule based system will be maintained. Machine learning will still be utilized but as a shadow mode and will not be enforced. This would build a significant amount of trust when the project continues. It is vital to carefully plan change management in the government system as it is highly regulated and adopting artificial intelligence-based detection systems requires additional attention to mitigate resistance and promote sustainability in using artificial intelligence. Major challenge addressed: As VA is a government agency where organizational rules and behavior are significantly important, there has been notable resistance to AI-based detection systems. The project experienced hesitance from senior stakeholders due to the lack of transparency in machine learning models and perceived disruption to existing workflows within the organization. Change Management Change management application:

Change management principles applied:

Back to Table of Contents