We ask AI to find meaning in our data — but the authentic story was gone before it ever looked
In What We Lost When Datum Became Data, I traced how a word that once meant "something given" shed its own etymology and became a synonym for "values in a system." Along the way, the who, the why, and the context quietly disappeared. I used the Groningen definition, the Employee-Person example, and the metadata stack to show where those losses happen.
That argument had its own reasons. It has sharper ones now.
Because we are now asking AI — Large Language Models in particular — to read the values we stored and behave as if the datum were still intact. It is not. And the models, being what they are, paper over the absence with confident invention. What gets called hallucination is, much of the time, an AI politely restoring the verbs, categories, and contexts we threw away — and restoring them wrong, because there was nothing left to restore them from.
So the question this piece tries to answer is narrower than the first: if the datum is what AI actually needs, what must we retroactively restore — and what does it take to get from data back to datum?
Download 
Use Online