In some redacted documents, there is even an alphabetical word index at the end with a list of pages on which the words appear.

The redacted words are also redacted in the word index, but the alphabetically preceding and succeeding words are visible, as is the number of index lines taken up by the redacted word's entry, which correlates with the number of appearances of that word.

This seems like rather useful information to constrain a search by such a tool.

I was thinking something similar. I wonder if the font uses kerning, and you know the rendering engine and the algorithm for how the text was blocked, if you can get exact text back even. Or, at a minimum, rule out words based on the available information. Not a field I am familiar with but I bet there are a lot of ways to uncover the redacted values.

This just attempts to match box dimensions.

Does it even matter? The kind of people who see stuff like this and are still fine with it are likely fine with anything else thats discovered as well.

The truth has become irrelevant.

https://www.justice.gov/epstein/files/DataSet%208/EFTA000250...

i'm sure people will ask chatGPT to do this very thing, so it's a good thing LLMs never make shit up