OpenAI is evaluating a technical solution to this problem by watermarking the output with small amounts of randomness, such that a series of tokens of output can be tested for having these small random variations.
Ask HN: How would you build a ChatGPT detector?
31–40 of 109 posts
Re: Ask HN: How would you build a ChatGPT detector?
#32According to ChatGPT: > 1. Start by training a machine learning model to identify features of natural language. You can use supervised learning techniques with a corpus of labeled data consisting of texts written by both humans and AI systems. > 2. Use natural language processing techniques to extract features from the text such as word frequencies, grammar, and syntax. > 3. Implement a scoring system to assign a num…
That's a very verbose way of saying TF-IDF + SVM.
Re: Ask HN: How would you build a ChatGPT detector?
#33Maybe detecting grammar and spelling mistake will give you a first hint
Re: Ask HN: How would you build a ChatGPT detector?
#34By no means is this foolproof, but the older the age of an account, the less likely it is to be a bot. Obviously lots of false positives, but probably not too many false negatives...
I am sure there will be a market in selling old accounts to time-wash bots.
Re: Ask HN: How would you build a ChatGPT detector?
#35Humans will become incentivized to sprinkle racial slurs and hate facts throughout their writing as proof of authenticity.
Re: Ask HN: How would you build a ChatGPT detector?
#36Earlier quoted context omitted.
I know this is fun for people but I have to say the GPT reply is getting old fast.
The default style is such, arguably by design, exactly so it’s easily detected and gets downvoted on forums like here. IMHO it’s an important feature, even if it wasn’t intended for this purpose.
It's a classic "start by restating the question" style which so often precedes fluffy, empty bullshit. It must be by design, and it's a pretty silly design. It doesn't seem intrinsically necessary.
Re: Ask HN: How would you build a ChatGPT detector?
#37Earlier quoted context omitted.
I suspect that in time, this will only accelerate the degree to which AI and human-authored text are indistinguishable from each other.
I've already sent text to customers 100% written by AI. Ethically dubious in a commercial setting perhaps, but higher quality text than I would be able to produce myself. I asked OpenAI and it said: It is not necessarily unethical to send customers text generated by AI, but it depends on the context and the specific situation. For example, if the text is being used to deceive or mislead customers, then it would be un…
People have been coping and pasting text into emails for millennia without explicitly telling the recipient. This shouldn’t be any different.
Re: Ask HN: How would you build a ChatGPT detector?
#38Voight-Kampff machine
Re: Ask HN: How would you build a ChatGPT detector?
#39I think you should begin with specifying the purpose of such a detector, apart from the obvious one that you want to know that you are talking to a real human being.
Re: Ask HN: How would you build a ChatGPT detector?
#40According to ChatGPT: > 1. Start by training a machine learning model to identify features of natural language. You can use supervised learning techniques with a corpus of labeled data consisting of texts written by both humans and AI systems. > 2. Use natural language processing techniques to extract features from the text such as word frequencies, grammar, and syntax. > 3. Implement a scoring system to assign a num…
I think you should begin with specifying the purpose of such a detector, apart from the obvious one that you want to know that you are talking to a real human being.