Canadian Public Data Licensing and Citizen Dividend Act

⚠ This is a flightplan, not a commitment. Flightplans are hypothetical policy proposals modelled in the Ducklings causal simulation. They become projects only when voted into adoption through Consensus. Metrics shown reflect simulation model outputs, not real-world outcomes.

$0.10B CAD

Fiscal Impact

Medium-term (2–5 years)

Time Horizon

Medium

Implementation Risk

Epoch 100

Continuum Entry

Overview

Establish mandatory licensing fees for commercial use of Canadian public datasets and citizen-generated data by AI developers above a revenue threshold, distributing 60% as a direct annual Citizen Data Dividend to every Canadian adult. The remaining 40% funds sovereign AI compute (D-04 CSAIA synergy) and data infrastructure. Makes data sovereignty tangible: every Canadian receives a cheque as their share of the value their data creates. Projected: $140-280 CAD/person/year at steady state. Alaska's Permanent Fund Dividend — but for the digital economy rather than oil.

Problem Statement

US companies train commercial AI models on Canadian government datasets (Statistics Canada, CMHC, Health Canada, CRA, NRCan), Canadian social media activity, Canadian healthcare research, and Canadian geospatial data — then sell the AI products back to Canadians at full commercial price. The estimated value of Canadian data used in foreign AI training is ~$2.5B/year. Canada collects $0 in licensing revenue. Canadian citizens receive $0 in return. The data_sovereignty_index of 28/100 reflects this: Canada has passed AI regulation (C-27/AIDA) without establishing a revenue claim on the asset being regulated. You cannot regulate what you do not price.

Proposed Approach

Tier 1 (AI revenue > $50M CAD): mandatory Canadian Data Royalty at 2.5% of annual Canadian revenue derived from products trained on Canadian data. Applies to model training, fine-tuning, and inference products where Canadian data constitutes >5% of training corpus. Tier 2 (AI revenue $10-50M): 1.2% royalty rate with 2-year phase-in. Canadian entities exempt from royalty (revenue stays in Canada). Canadian Data Dividend Trust: receives all royalty revenue; distributes 60% equally to all Canadian adults (permanent residents included after 3 years) in annual payment; 25% to D-04 CSAIA sovereign compute investment; 15% to Statistics Canada data infrastructure. Enforcement: CRA collection mechanism; AIDA compliance integration.

Anticipated Impacts

foreign_ai_training_data_extraction_value falls from $2.5B as licensing displaces extraction; canadian_data_royalty_revenue grows from $0 to $0.8-1.4B/year by Year 3; citizen_data_dividend_per_capita rises from $0 to $140-280/person/year; ai_data_licensing_fund_balance accumulates to $2-4B by Year 5; data_sovereignty_index rises from 28 toward 45-55 (combined with D-04); sovereign_ai_compute_capacity rises as 25% of fund flows to D-04 CSAIA capital.

Ducklings Simulation

This proposal is active in the Ducklings causal simulation (Epoch 100). The simulation models downstream effects using a BFS cascade engine with strength-weighted, time-delayed edges capped at 3-hop depth and ±25% per-hop limits. Cascade outputs are bounded by variable saturation thresholds.

Domain: Economic Policy | Proposal ID: 196 | Series: E-series

How to Engage

Discuss this flightplan in the Pond forum under Economic Policy. Vote on adoption through Consensus. Adopted flightplans become projects with real-world implementation tracking.

Contact: [email protected]