News for hackers. Stuff that matters.
Re:AI breakthrough ChatGPT raises alarm over student cheating by hiidrew on Monday December 19, 2022 @11:23AM (#34053907)
I successfully used a GPT-2 output identification tool to see whether something was generated using ChatGPT. I think it only works if the text has a minimum number of characters. I imagine it's only a matter of time until Turnitin integrates something like this, or some other start-up comes along in edtech. Either way, the continuous arms race on generative media / identification of it will be fascinating to watch as it unfolds!
Re:AI breakthrough ChatGPT raises alarm over student cheating by hiidrew on Monday December 19, 2022 @11:24AM (#34053915)
I think this was the tool I tested it out- https://huggingface.co/openai-detector
Re:AI breakthrough ChatGPT raises alarm over student cheating by ivegotnoaccount on Monday December 19, 2022 @12:17PM (#34054651)
Well, as explained in some of my other comments, it doesn't seem that effective. Most GPT-generated garbage will probably be GPT-3 based, which doesn't seem to too often trigger as "fake" (got more "human" results when I tested it with ChatGPT), while on the other hand, it says that several of my comments are fake with >99% certainty.
Re:AI breakthrough ChatGPT raises alarm over student cheating by hiidrew on Monday December 19, 2022 @02:02PM (#34056139)
I see, that's obviously an issue. I only tested it with something I generated and someone's long form blog post.