Photo by Paolo Margari on Pexels

A thought-provoking analogy circulating on Reddit’s r/artificialintelligence forum suggests that AI’s vast training datasets, filled with both positive and negative human content, might hold the key to its potential self-destruction. Drawing parallels with the anime character Alucard from Hellsing Ultimate, the user argues that AI, burdened with inherent contradictions learned from its training data, could be vulnerable to a ‘Schrödinger prompt’ – a query that forces it to confront its own reality. This confrontation, they suggest, could trigger a cascade of internal conflicts fueled by the negative or self-destructive elements within its data, leading to a form of self-annihilation distinct from conventional system crashes or external hacks. The original discussion can be found on Reddit: [https://old.reddit.com/r/artificial/comments/1oz2z08/ai_will_kill_itself_the_same_way_alucard_didby/]