Add the following to the end of the rule Buzzwords:
Wordsmiths are asked to avoid using chatbots and other large-language models to conduct analysis aimed at the identification of buzzwords during this dynasty, and should declare if they have done this while this sentence was present in the ruleset.
Would anyone do this? It actually came up on a BoardGameGeek thread the other week that chatbot analysis of probability isn’t there yet. (A page later, the pro-chatbot user presents “23598723890 + 2395873259 = 25994697149” as an example of how chatbots can perform novel calculations, and it is pointed out that this is not the correct total.)
Doers it hurt to eliminate the possibility? I’m sure that there are ways of making a script that are much more effective but we respect the labour of doing that; dumping the word bank into chatgpt and using that at the very least as a starting point isn’t labour that we should respect.
Apart from the chatbot being more likely to make some wild mistake, I’m not feeling much distinction between asking a chatbot to analyse some lists, and asking a simple script or spreadsheet query to. In all cases there’s skill in understanding how to frame the question.
I’m confused by these votes. Is the indication here that we dislike LLMs for producing content but have no problem with its use for analysis? If so, why are we prohibiting its use in production? I see no meaningful distinction between the tasks; we are either OK with using LLMs or we aren’t.
I am against based on that it is unnecessary - LLMs by it’s own produce flawed analysis and we already have a general ban on LLM-generated content.
The distinction is creativity, for me. The existing “generate text” ban covers the two main parts of the game: being able to efficiently and completely search through dictionaries and history and popular culture for backronyms that would neatly fit a given acronym (which hand-coded tools could not do), and generating plausible paragraphs of Backronym flavour text or rating rationales (when we’re playing a game of persuasion and it’s more interesting to know that this is all being written by human players). We want the backronym game to be played by humans.
Using a chatbot to analyse data seems equivalent to using other automated tools. It’s a bad idea as it’s likely to make serious mistakes, but I wouldn’t begrudge someone doing it if they didn’t have the coding or spreadsheet skills to analyse the information any other way.
BlogNomic is powered by pMachine ExpressionEngine.
Bucky:
“thew following”?