Earlier quoted context omitted.
[flagged]
It's your fault that you deny the widely accepted Babi Yar massacre and call it "a Russian fake", exactly what Holocaust deniers do. If the shoe fits... So in your opinion, countries such as all of Europe (Germany included), the US, Israel, etc are all part of this conspiracy?
We need to tell people ChatGPT will lie to them, not debate linguistics
431–440 of 485 posts
Re: We need to tell people ChatGPT will lie to them, not debate linguistics
#432Earlier quoted context omitted.
chatgpt doesn't ignore the information you put into its prompt window. but also giving it the complete context does not guarantee you correct information, because that's not how it works. Your earlier comment was "You could just ask ChatGPT what years the Mets won and it will tell you the correct answer." -- that is not accurate. That's my point. It doesn't know those facts. You might ask it that one minute and get a…
A series of claims have been made that can be objectively tested for their veracity, namely that pasting information, whether from Wikipedia or another source, is a way of ensuring that ChatGPT does not make false claims, or as OP stated it, ChatGPT will not "synthesize" a completion when provided with an augmented prompt that contains a sufficiently large context (such as a Wikipedia article). I have conducted such…
Here is my research: https://github.com/williamcotton/empirical-philosophy/blob/m...
It is clear that analytic augmentations will result in more factual information.
Your claims are unfounded and untested.
Re: We need to tell people ChatGPT will lie to them, not debate linguistics
#433Earlier quoted context omitted.
[flagged]
It's your fault that you deny the widely accepted Babi Yar massacre and call it "a Russian fake", exactly what Holocaust deniers do. If the shoe fits... So in your opinion, countries such as all of Europe (Germany included), the US, Israel, etc are all part of this conspiracy?
"I know that many Jews were shot by Germans in the Rivne region."
"Yes. They are buried in the Sosenki area. We can drive there in a car. Would you like me to introduce you to relatives who can tell you more about it?"
"No, that's not necessary. Where are the Kiev Jews who were shot in Babi Yar buried?"
"I don't know. Isn't your mother from Kiev? You should know where your relatives are buried."
"I don't know anything about them. I'll ask my mother."
Later on:
"Mom, do you know anyone from our acquaintances who was shot in Babi Yar?"
"The Germans shot Ukrainians in Babi Yar, but it was forbidden to talk about it in the Soviet Union, just like with the Holodomor, or else you would be arrested."
I am a descendant of Jews who lived in Kiev during World War II. But I know nothing about the Jews who were shot in Babi Yar. My relatives or acquaintances do not know either. My search led me to the grave of Ukrainians who were shot in Babi Yar and buried near the radio mast. This grave is visible in aerial reconnaissance photographs.
If you know who was shot and where they are buried, please tell me. I would be grateful. We can erect a monument at this location.
Re: We need to tell people ChatGPT will lie to them, not debate linguistics
#434Earlier quoted context omitted.
I agree that lying also has a connotation of intent and therefore intelligence, and so is at least technically inappropriate. Yet I agree with the researchers that the benefits of using "lying" instead of the more technically accurate "confabulation" outweigh the negatives. In communicating with new users who may over-rely on it, getting across highly technically correct metaphors is less important than delivering th…
What’s wrong with: “it is often wrong”? I - or GPT - can think of various variations that do not have links with “lying”. I think it is a bad idea. Let people use it and see for themselves. It has unique failure modes, not unlike every other tool we have. You will only learn by interacting with it. Some people are in the habit of “protecting the people” but I am not quite sure that is necessary.
Good q; the problem is the passive voice
The passive voice connotes far less urgency and importance to any warning. Even "It is often dangerously wrong.", is still a more passive phrasing than "It will lie to you, with a highly confidant tone."
I also cringe a bit at using active verbs that imply agency and intent in a statistical model; it makes an inaccurate implication about the technology. But the same applies whether we call it "lying" or "confabulating".
I also do not see it as 'paternalistically protecting the user', but simply as 'truth in advertising': - accurately representing the strengths and weaknesses of the product.
I absolutely agree that everyone should try it themselves - there are MANY useful use-cases. But they should simply be accurately forewarned, in no uncertain terms, about it's behavior.
Re: We need to tell people ChatGPT will lie to them, not debate linguistics
#435The part that's concerning about ChatGPT is that a computer program that is "confidently wrong" is basically indistinguishable from what dumb people think smart people are like. This means people are going to believe ChatGPT's lies unless they are repeatedly told not to trust it just like they believe the lies of individuals whose intelligence is roughly equivalent to ChatGPT's. Based on my understanding of the appro…
> The only danger is that stupid people might get their brains programmed by AI rather than by demagogues which should have little practical difference. This may be the best point that you've made. We're already drowning in propaganda and bullshit created by humans, so adding propaganda and bullshit created by AI to the mix may just be a substitution rather than any tectonic change.
I know more academic/intellectual types who are less willing to, than I do the average joe who seeks answers from all directions and discerns accordingly.
Re: We need to tell people ChatGPT will lie to them, not debate linguistics
#436Earlier quoted context omitted.
It's your fault that you deny the widely accepted Babi Yar massacre and call it "a Russian fake", exactly what Holocaust deniers do. If the shoe fits... So in your opinion, countries such as all of Europe (Germany included), the US, Israel, etc are all part of this conspiracy?
Also weird how someone supposedly so deeply informed has somehow missed both the existing memorial and the year-long fights over building additional ones and has no clue where they might be placed?
Furthermore, the Security Service of Ukraine warns[0] us that this memorial is being built by Russians in order to discredit Ukraine. And we completely agree with them.
[0]: https://censor.net/ua/news/3243965/sbu_informuvala_shmygalya...
Re: We need to tell people ChatGPT will lie to them, not debate linguistics
#437Earlier quoted context omitted.
The problem is that it will be cheaper at scale. That will allow the BS to be even more targeted growing the population of acolytes of ignorance. I don't know how much bigger it is, but it seems like we're around 20% today. If it gets to a majority there could be real problems.
There's already more information generated each day than any individual can ever hope to process, adding more crap to that pile doesn't move the needle much if at all.
Re: We need to tell people ChatGPT will lie to them, not debate linguistics
#438Earlier quoted context omitted.
> That's lying. If you don't know the answer, the correct answer is "I don't know", or to ask clarifying questions. To pretend that you know when you don't, is to lie. You can use that definition, but it is not how the word "lie" is commonly used. Mirriam-Webster defines lie as "to make an untrue statement with the intent to deceive"[0] Cambridge dictionary defines lie as "to say or write something that is not true i…
> You can use that definition, but it is not how the word "lie" is commonly used. The problem is that any word that ascribes agency to the LLM will technically be incorrect. But that removes most possible descriptions of its tone and style which are a relevant part of its response. An analogy would be if ChatGPT started insulting me and calling my question stupid. Would it be wrong to call its response "rude" or "mea…
It’s important because of motive. A lie is told with intent to conceal a known truth. It’s not just LLM agency that’s in question, it’s human malice.
Co-opting the term “lie” is a rhetorical tool used to shift the conversation from “this person/LLM is saying things that aren’t true” to “this person/LLM is acting maliciously and needs to be punished/sanctioned”
All I want is an honest conversation. It’s more than a little ironic that many humans are themselves being less than honest (and sometimes just bullshitting!) about the context when LLMs produce false statements.
Re: We need to tell people ChatGPT will lie to them, not debate linguistics
#439Earlier quoted context omitted.
On the subject of not knowing thigs... Your clain is incorrect. Prompt: Tell me what you know about the Portland metro bombing terrorist attack of 2005. GPT4: I'm sorry, but I cannot provide information on the Portland metro bombing terrorist attack of 2005, as there is no historical record of such an event occurring. It's possible you may have confused this with another event or have incorrect information. Please pr…
There is a big difference between what they are talking about (I believe) and asking it factual questions withal straightforward answers. Try discussing something a bit more complex and it’s very easy to get ChatGPT to start contradicting itself all the time (which is more or less the same thing). From my experience ChatGPT is prone to come up with lists of “probable” answer when it’s not feeling very confident. Then…
Re: We need to tell people ChatGPT will lie to them, not debate linguistics
#440Earlier quoted context omitted.
ChatGPT no more cares to impress or persuade you as it cares to tell the truth or lie. It will say what its training maps to its model the best. No more, no less. If you believe it or not -- ChatGPT doesn't care -- except to the extent that you report the answer and they tweak its training/model in the future.
It definitely cares about impressing and persuasion. Those are things that will result in higher ratings from the user