What happens when you put AI to work deciphering lost languages?

AI is being applied to some of history's most stubborn linguistic puzzles, including Linear A (the undeciphered writing of Minoan Crete) and Etruscan. A recent case highlights both the promise and limits: a self-taught engineer used AI-assisted scripts to test a Semitic-root hypothesis against the full Linear A corpus, assigning values to 40 signs and building a 408-word lexicon. The key takeaway is that AI excels at pattern-matching at scale, but the creative leap — the initial hypothesis — still came from a human. AI amplifies human intuition rather than replacing it.