Comment by yunseo47
2 days ago
Whether it’s code, general text, or university assignments, the core issue is taking responsibility for one's own work. While I share the concerns raised in this thread, I believe the focus on 'LLM usage' is a bit of a red herring. The fundamental principle of ownership hasn't changed with the advent of LLMs; the tool itself isn't the issue, but rather the abdication of responsibility by the author.
For instance, if a non-native speaker translates their own writing using machine translation or an AI, is that problematic—provided they personally review and vet the content before posting? I don't think the people calling out AI use on this board are taking issue with that. Ultimately, it’s not about the method; it’s about the author's attitude.
The reason LLMs are so disruptive now is that while "shitposts" used to be obvious, we're now seeing "plausible" low-effort content generated without any human oversight. Irresponsible people have always been around, but LLMs have given them the tools to scale that irresponsibility to an unprecedented level.
I think a human-like piece with minor mistakes resonates more emotionally than a perfectly written piece that looks like it was written by AI. However, since there seems to be a grammar debate going on here, I'd like to add: Is it a bad thing for non-native speakers to use AI to correct grammar or awkward expressions? I think it definitely has positive aspects in terms of lowering language barriers.
> the tool itself isn't the issue, but rather the abdication of responsibility by the author
The biggest current social problem with AI content is our collective lack of transparency into how much human responsibility was taken.
Give a <100% reliable/accurate AI tool, the same post/code may have had {every line vetted by a human} or {no lines vetted by a human}... and readers have no way of telling which it is!
Because even if no edits needed to be made, the former carries a lot more signal than the latter, because it reduces risk of AI slop and therefore makes the content more valuable.
At the same time, it also costs more time to produce, so in any competitive marketplace (YouTube, paid comments, startup code, etc.) the unvetted AI content will dominate.