Consent Shifts Threaten Data Moat
General Intuition
The key risk is that this data advantage is partly a permissions advantage, not just a technology advantage. If users, regulators, or platform partners become less comfortable with gameplay clips being reused for AI training, the usable dataset can shrink even while the models stay just as good and rivals stay in the same place. That matters because the moat comes from continued access to fresh, action labeled clips at massive scale, not from a one time archive alone.
-
Medal’s privacy policy says clips, metadata, and derived data can be shared for research, AI development, and model training, and it also says users can opt out of sharing personal information with third parties in app privacy settings. That means the training pipeline depends on an ongoing consent structure, not only on technical collection capability.
-
The joint controller setup raises the stakes because data rights are tied to a shared governance framework. Under GDPR guidance, joint controllers share responsibility for how data is processed and how user rights are handled, so any tightening in consent standards or enforcement can directly affect what data remains usable for model training.
-
General Intuition’s own positioning makes the dependency concrete. The company is built around Medal’s billions of gameplay clips, and Medal has publicly described those clips as the raw material for spatial intelligence products. If social expectations shift toward private by default capture, narrower reuse, or more visible opt outs, the moat weakens even if no competitor builds a better model.
Going forward, the strongest data moats in AI will belong to companies that pair large scale collection with durable user permissioning. In this market, defensibility will come less from having clips today, and more from preserving the legal and social right to keep turning tomorrow’s clips into training data at scale.