TerpInsights

AI Lab

Ethical AI Infrastructure for Regulated Markets

TerpInsights AI Lab is a research-grade applied AI testbed designed to evaluate transparency, bias, and robustness in forecasting systems deployed within regulated public markets.

Built in Maryland. Designed for broader institutional collaboration.

Explore the Framework Request Collaboration Discussion

Why an AI Lab?

AI systems increasingly influence public markets, regulatory decisions, and economic opportunity. Yet most forecasting tools operate as black boxes — with limited documentation, minimal bias testing, and no structured stress evaluation.

The AI Lab exists to change that.

This initiative transforms Maryland's regulated cannabis market into a live validation environment for:

  • •Interpretable forecasting
  • •Segment-level bias detection
  • •Policy shock simulation
  • •Equity participation measurement
  • •Transparent model documentation

The goal is not just prediction accuracy — but accountable deployment.

Core Modules (Preview)

Structured frameworks for interpretability, fairness, and robustness.

1

Forecast Lab

Interpretable revenue and volume forecasting with documented accuracy metrics, confidence intervals, and feature explainability.

Includes:

  • • MAE, MAPE, RMSE, R² reporting
  • • Time-series cross validation
  • • Feature importance diagnostics
  • • Automated retraining triggers
2

Fairness Audit

Segment-level error analysis and parity diagnostics to detect underprediction or overprediction across equity-sensitive populations.

Includes:

  • • Error gap tracking
  • • Parity threshold flags
  • • Sample size confidence warnings
  • • Remediation protocol documentation
3

Stress Testing

Simulated regulatory and market shocks to assess model robustness under policy changes.

Includes:

  • • Tax change scenarios
  • • Policy event mapping
  • • Revenue impact modeling
  • • Robustness scoring
4

Equity & Access Metrics

Quantitative indicators measuring market participation, opportunity distribution, and disparity trends.

Includes:

  • • Revenue share by equity status
  • • Regional participation tracking
  • • Experimental Opportunity Index (under review)
5

Data & Method Transparency

Full documentation of:

  • • Data sources
  • • Feature engineering
  • • Model architecture
  • • Evaluation methodology
  • • Known limitations

No black boxes.

Research Status

Current Status: Active Development (v1)

The AI Lab is currently operating as a structured prototype.

  • ✓ Forecast models deployed and monitored monthly
  • ✓ Fairness testing framework v1 implemented
  • ✓ Policy NLP module under development
  • ✓ Methodology documentation open for institutional review

We are actively exploring academic collaboration to refine fairness metrics, expand intersectional analysis, and formalize validation standards.

Who This Is For

The AI Lab is designed for:

AI ethics researchers Public policy scholars Regulatory agencies Economic development organizations Civic data practitioners

If you are studying responsible AI deployment in real-world systems, this environment is intended to serve as a collaborative validation space.

Partnership & Research Inquiry

TerpInsights is exploring collaboration with universities and research centers focused on:

  • • Bias detection & remediation
  • • Model stress testing frameworks
  • • AI governance standards
  • • Equity measurement methodologies

If you are interested in research partnership or structured evaluation, we welcome conversation.

Request Collaboration Discussion

TerpInsights AI Lab is an applied civic AI infrastructure initiative focused on transparency, fairness, and accountable deployment in regulated markets.