Post AyMLW8urstmc1bXzl2 by tommorris@mastodon.social
 (DIR) More posts by tommorris@mastodon.social
 (DIR) Post #AyMLW8urstmc1bXzl2 by tommorris@mastodon.social
       2025-09-19T08:18:26Z
       
       0 likes, 0 repeats
       
       If I trained a large language model with the goal of making evidence-free libel, but with no specific goal to libel any particular person, am I liable?As in, it takes data from the web, subtly misinterprets it, conflates distinct people, and treats hazy/unevidenced connections as certainties… and I’m reasonably confident it will make mistakes quite frequently, but I have no idea who it will hurt.Basically imagine automating Redditors investigating people after crimes or conspiracy logic.
       
 (DIR) Post #AyMLWAIejwPCJfgVTU by cnx@awkward.place
       2025-09-19T09:13:33.088814Z
       
       0 likes, 0 repeats
       
       If I trained a large language model [...], am I liable?  As in, it takes data from the webYes, @tommorris, unless thou art a corporation with billions to bribe lobby the government, thou would likely get a life sentence from copyright infringement.