←back to thread

427 points JumpCrisscross | 1 comments | | HN request time: 0s | source
Show context
greatartiste ◴[] No.41901335[source]
For a human who deals with student work or reads job applications spotting AI generated work quickly becomes trivially easy. Text seems to use the same general framework (although words are swapped around) also we see what I call 'word of the week' where whichever 'AI' engine seems to get hung up on a particular English word which is often an unusual one and uses it at every opportunity. It isn't long before you realise that the adage that this is just autocomplete on steroids is true.

However programming a computer to do this isn't easy. In a previous job I had dealing with plagiarism detectors and soon realised how garbage they were (and also how easily fooled they are - but that is another story). The staff soon realised what garbage these tools are so if a student accused of plagiarism decided to argue back then the accusation would be quietly dropped.

replies(14): >>41901440 #>>41901484 #>>41901662 #>>41901851 #>>41901926 #>>41901937 #>>41902038 #>>41902121 #>>41902132 #>>41902248 #>>41902627 #>>41902658 #>>41903988 #>>41906183 #
acchow ◴[] No.41901484[source]
> For a human who deals with student work or reads job applications spotting AI generated work quickly becomes trivially easy. Text seems to use the same general framework (although words are swapped around) also we see what I call 'word of the week'

Easy to catch people that aren't trying in the slightest not to get caught, right? I could instead feed a corpus of my own writing to ChatGPT and ask it to write in my style.

replies(1): >>41901583 #
hau ◴[] No.41901583[source]
I don't believe it's possible at all if any effort is made beyond prompting chat-like interfaces to "generate X". Given a hand crafted corpus of text even current llms could produce perfect style transfer for a generated continuation. If someone believes it's trivially easy to detect, then they absolutely have no idea what they are dealing with.

I assume most people would make least amount of effort and simply prompt chat interface to produce some text, such text is rather detectable. I would like to see some experiments even for this type of detection though.

replies(1): >>41901673 #
hnlmorg ◴[] No.41901673[source]
Are you then plagiarising if the LLM is just regurgitating stuff you’d personally written?

The point of these detectors is to spot stuff the students didn’t research and write themselves. But if the corpus is your own written material then you’ve already done the work yourself.

replies(2): >>41901696 #>>41901754 #
1. hau ◴[] No.41901754{3}[source]
Oh I agree, producing text by llms which is expected to be produced by human is at least deceiving and probably plagiarising. It's also skipping some important work, if we're talking about some person trying to detect it at all, usually in education context.

Student don't have to perform research or study for the given task, they need to acquire an example of text suitable for reproducing their style, text structure, to create an impression of being produced by hand, so the original task could be avoided. You have to have at least one corpus of your own work for this to work, or an adequate substitute. And you still could reject works by their content, but we are specifically talking about llm smell.

I was talking about the task of detecting llm generated text which is incredibly hard if any effort is made, while some people have an impression that it's trivially easy. It leads to unfair outcomes while giving false confidence to e.g. teachers that llms are adequately accounted for.