AI-Powered Proctoring Engine Development & Transparency

Texas A&M University is seeking a subcontractor to develop and maintain AI models designed to detect suspicious behavior during academic exams, with a strong emphasis on ethical deployment and operational transparency. The contract requires the contractor to build systems that identify potential academic dishonesty through behavioral analysis, while ensuring complete clarity in how the models function, what data they use, and how decisions are made. Importantly, the use of any university data for training or refining these models is strictly prohibited, requiring the contractor to rely solely on externally sourced, anonymized, or synthetic datasets to ensure privacy and compliance with institutional policies. The contractor must prioritize transparency throughout the development lifecycle, documenting all model architecture, training methodologies, and performance metrics to enable independent review and accountability. The system must be designed to avoid bias, ensure fairness across diverse student populations, and provide explainable outputs that can be audited by university personnel. All work must be completed under the terms of the subcontract, with final deliverables due by July 8, 2026, and the solution must be deployable in the university’s existing exam infrastructure without compromising system integrity or student data privacy.