Do you think anyone can use this G-OpenAI platform to write posts in bitcointalk forum, and is there a way to detect this?
Maybe we need to have some kind of detector for text written by bots, but I often think that some people are worst than AI after I read their posts
You would probably need to insert the thread history as a contextual basis if you wanted consistency. I mentioned use of GLTR in a previous thread.Maybe we need to have some kind of detector for text written by bots, but I often think that some people are worst than AI after I read their posts
You can tell from the prose whether a writer is human or not. Though, from some of the higher-end side, it is quite pleasant to read. Maybe it's low-effort, but if someone were to generate and pick out interesting AI-generated content as a substitute to writing compelling responses (i.e. prompt-writing vs. exhaustive creation) that would be far better than most things posted out there. An interesting question to raise is whether it could be construed as plagiarism. And subsequently, what changes would be necessary to transform it into an independent derivation, if any at all?