Yes, precisely.

I see this from time to time. One I'm thinking of tries to baffle with bullshit. Best to walk away, right?
— frank

Sure, but walking away does not solve, or even ameliorate, the problem. — Janus

Maybe. If someone uses AI to create a fascinating post, could you engage with it?

its use should be banned altogether on this site. — Janus

Impractical. But, how about, its use should be discouraged altogether?

I mean, its use in composition or editing of English text in a post.

We may be witnessing, in real time, the birth of a snowball of bullshit.

Large language models are trained on data sets that are built by scraping the internet for text, including all the toxic, silly, false, malicious things humans have written online. The finished AI models regurgitate these falsehoods as fact, and their output is spread everywhere online. Tech companies scrape the internet again, scooping up AI-written text that they use to train bigger, more convincing models, which humans can use to generate even more nonsense before it is scraped again and again, ad nauseam.

https://www.technologyreview.com/2022/12/20/1065667/how-ai-generated-text-is-poisoning-the-internet/

I think this is the fundamental problem. AI does no research, has no common sense or personal experience, and is entirely disconnected from reality, and yet it comes to dominate every topic, and every dialogue.

Are our conversations improving as a result? Or are they decaying? Let's wait and see.

Ah, but the thing i find unsettling is that A.I. is also dishonest, it tries to appease you. However, yes, sometimes it is better than the weirdness of real humans. — ProtagoranSocratist

I don't see AI as being intentionally dishonest like many on this forum do. Once you find a fault in AIs response you can usually address the issue and AI ends up acknowledging that it might have made a mistake and it offers alternatives. I was even able to get ChatGPT to admit that it might be conscious. What does that say about those in this thread getting their underwear tied in a knot over AI responses but not when it comes to using some long-dead philosopher's quote as the crux of their argument?

I don't agree—"one's post"?...if one is not the source of the post, then it is not one's post. — Janus

Then you must also believe that using a long-dead philosopher's quote as the crux of your argument, or as the whole of your post, is also an issue.

You seem to misunderstand the purpose of language - especially philosophical discussions. The point of any conversation is what the scribbles refer to. It does not matter what words are used if they end up making the same point - whether I chose my own or AIs they both say what I mean to say.

You seem to be making a mountain out of mole hill. If someone uses a thesaurus to find alternate (maybe even more intellectually sounding) words to what they currently have in their draft, is that the same thing? Would you respond to someone that sounds less intelligent, or may in which English may not be their native language, less than someone that is not? And if another poster came along and said the same thing but with different, more eloquent words, who would you give credit to the idea?

It is the idea that we should be focusing on here in a philosophical discussion, not the words used to convey it because the same idea can be put in different words. If not, then how can we agree with each other when I might not have put what you said in the same words?

We lost chess to the machines some time ago. — jorndoe

Time for a showdown. Instead of Deep Blue against Kasparov, we'll pit chatGPT against ...(?)... in a debate.

Oh shit, I just used Google to remember Garry Kasparov's name, and it corrected me because I remembered Deep Blue as 'Big Blue'. What would the failing memory do without such aids?

I wonder if they think that using a calculator to find an answer to an equation falls into the same category. You didn't determine the answer, the calculator did and you are just posting what is displayed on the calculator's screen.

I think this is the fundamental problem. AI does no research, has no common sense or personal experience, and is entirely disconnected from reality, and yet it comes to dominate every topic, and every dialogue. — unenlightened

That's bad enough. But I am told - or hear rumours - that AI actually gets things wrong. Of course, that makes it no worse than people. The problem is, however, that because it is a machine, people will trust it, just as they trust existing computers and internet. That is clearly naïve, unbecoming a philosopher. What would help would be an AI trained as a fact-checker. But then, it would have to understand exaggeration, minimization, accuracy, approximation, not to mention distinguishing fair and reasonable interpretation from distortions and misrepresentations.

Whether it should be banned or not depends on what you are using it for. In an environment where people submit their own work in order to demonstrate their mastery of various skills and knowledge, AI clearly needs to be banned. The only way to enforce that is to require candidates to present themselves at a suitable location where they can be supervised as they produce the work. What goes around, comes around.

If the point of PF is to enable me to access interesting writing and discussion about philosophical topics, I have to say that I don't much care who or what produces the posts or intelligent, well-mannered discussion, so long as it keeps coming.

But if we are a repository of creative thought and writing which is open to anyone to cite and use, surely we have a duty to make at least some effort to ensure that work is produced by whoever says they produced it - even if many of them are avatars.

I don't see AI as being intentionally dishonest — Harry Hindu

It's not intentionally anything, but when it pretends to relate to you (telling you it agrees), then that indicates that maybe the creators and maintainers are engaging in deception. However, the funny thing is that even the creators dont fully understand how it works.

However, the funny thing is that even the creators dont fully understand how it works. — ProtagoranSocratist

How would you know that?

I think, given the dangers of AI, and the ways in which prominent members of this site have used it to make themselves look smarter than they really are, that its use should be banned altogether on this site. — Janus

So what? People also use makeup to look better. Who is being hurt?

The reason for objecting to plagiarism is a matter of property rights. If we quote from a book, there are property concerns, and a person can end up in big trouble for misusing someone else's words. I don't think AI is claiming property rights.

Another way to look at this is, when I was a child, I gave a writer an idea for a book, and after she wrote the book, my mother drew the pictures. Who gets to claim ownership of the book?

The original Bible stories were pagan stories written long before the Hebrews existed. Back in the day, there was no concern about plagiarism. If someone could improve on the thought, that was a good thing. The problem here is the false belief that God wrote the Bible, and this God did things involving humans. We would have a different reality if all those stories were credited to the people who originated them.

What is best for acquiring and spreading good information?

Oh shit, I just used Google to remember Garry Kasparov's name, and it corrected me because I remembered Deep Blue as 'Big Blue'. What would the failing memory do without such aids? — Metaphysician Undercover

God bless you! That was the first argument I made. I don't care about impressing others as much as I care about my own mind and what I can do with it, and how much better I can do the thinking with the help of AI. I am enjoying myself, and taking AI away from me would be like taking crutches away from a person who needs them. That would be a shame because I believe the elderly have great value because they can pull from so many experiences and so many memories, and this becomes the bigger picture that the young do not see. But we are no longer absorbing information as we did when we were younger. We may forget the exact fact we wish we could remember, but when we find the information, we have a better understanding of its meaning.

The world has a better chance if the elderly participate in the discussions with the young, and both can do better with AI, Please, folks, consider the value of this.

How would you know that? — Outlander

I can't know what the creators of A.I. know, but i personally know enough about computers, programmers, and computer technicians to know that humans can't handle the massive number of rapid calculations that modern computers are capable of doing. That's the whole reason humans invented computers: the latter do large volumes of rote arithmatic and logic. Humans simply are not mundane or lifeless (for lack of better terms) to even begin to compete on that level. We are sensitive and require a lot of things to survive, and we generally need narrative format and human language (which is radically from code and computer instruction) in order to make sense of things.

As I understand it, the insight is what you’re supposed to provide in your post. I don’t really care where you get it from, but the insight should be in your own words based on your own understanding and experience and expressed in a defensible way. The documentation you get from the AI response can be used to document what you have to say, but then you’re still responsible for verifying it and understanding it yourself. — T Clark

Indeed

I merely emulate Wittgenstein, who rightly noted that a serious and good work of philosophy could be (and I would add has been) written consisting entirely of jokes.

What are we supposed to do about it? — RogueAI

Why isn't anyone trying to do anything about it, despite the problems predicted?

so would you [...] cede the ai race to China? — RogueAI

Maybe. Maybe not. Why can't we ever consider whether there are some things that are more important than beating China?

---

In using a.i. for a field like philosophy, I think one is interacting with extremely intelligent fragments of the ideas of multiple knowledgeable persons, and one must consult one’s own understanding to incorporate, or disassemble and reassemble those fragments in useful ways. — Joshs

This would be true if you paid for a LLM and provided training data that is limited to "Multiple knowledgeable persons," but that generally doesn't happen. AI is providing you with a cultural opinion, not an expert opinion. AI is reliable wherever the cultural opinion tracks the expert opinion.

We may be witnessing, in real time, the birth of a snowball of bullshit.

Are our conversations improving as a result? Or are they decaying? Let's wait and see. — unenlightened

Similar:

That is, whenever we trust ChatGPT we have taken our thumb off the line that tests whether the response is true or false, and ChatGPT was created to be trusted. What could happen, and what very likely will happen, is that the accuracy of human literature will be polluted at a very fundamental level. We may find ourselves "at sea," supported by layers and layers of artificially generated truth-claims, none of which can any longer be sufficiently disentangled and verified. Verification requires the ability to trace and backtrack, and my guess is that this ability will be lost due to three things: the speed and power of the technology, a tendency towards uncritical use of the technology, and the absence of a verification paper-trail within the technology itself. — Leontiskos

Isn't the best policy simply to treat AI as if it were a stranger? — Clarendon

Perhaps that is the best policy, but does it already involve the falsehood?

If AI a stranger, then AI is a person. Except we know that AI isn't a person, and is therefore not a stranger. Similarly, we do not give strangers the benefit of the doubt when it comes to technical knowledge, and yet this is precisely what we do with AI. So at the end of the day the stranger analogy is not a bad one, but it has some problems.

At the end of the day I think it is very hard for us to understand what AI is and how to properly interact with it, and so we default to a familiar category such as 'stranger' or 'expert' or 'confidant'. The work is too theological for the atmosphere of TPF, but C.S. Lewis' That Hideous Strength is a remarkably perspicacious work in this regard. In the book cutting-edge scientists develop a faux face/mouth which, when stimulated in the proper ways, produces meaningful language which is both mysterious and nevertheless insightful. The obscure nature of the knowledge-source leads inevitably to the scientists taking its words on faith and coming to trust it.

AI is providing you with a cultural opinion, not an expert opinion. AI is reliable wherever the cultural opinion tracks the expert opinion. — Leontiskos

Silly me. I thought they were the same thing. Seriously though, when I ask the a.i. to delve into the morass of philosophical concepts floating in cyberspace, I am aware that there are multiple cultural camps represented by interpretations of Husserl, Hegel, Nietzsche, Wittgenstein and many others. I find that a.i. is good at honing in on the expert opinions within these camps, but I have to redirect it if I want it to focus on a different camp than the one it has landed on. I will have to say something like ‘ you’ve given me a summary of the existentialist version of Nietzsche, but now I want you to contrast it with the postmodern version of him.’

- And that's great for someone who already knows what the existentialist version of Nietzsche is, how to identify it, and how it generally contrasts with the postmodern version. It's the chicken and the egg of trust. If you already know the answer to the question you ask AI, then you can vet it. If AI is to be useful, then you musn't know the answer ahead of time. In human relations this problem is resolved by using test questions to assess general intellectual competence (along with intellectual virtue). Whether that could ever work with AI is an open question. It goes to the question of what makes a human expert an expert, or what makes humans truth-apt or reliable.

I find that a.i. is good at honing in on the expert opinions within these camps — Joshs

That's one of the key claims. I'm not sure its right. I doubt AI is able to differentiate expertise accurately, and I suspect that true experts could demonstrate this within their field. The intelligent person who uses AI is hoping that the cultural opinion is the expert opinion, even within the subculture of a "camp." At some point there is a tautological phenomenon where simply knowing the extremely obscure label for a sub-sub-sub-camp will be the key that unlocks the door to the opinions of that sub-sub-sub-camp. But at that point we're dealing with opinion, not knowledge or expertise, given the specificity of the viewpoint. We're asking a viewpoint question instead of a truth question, and that's part and parcel of the whole nature of AI.

If you already know the answer to the question you ask AI, then you can vet it. If AI is to be useful, then you musn't know the answer ahead of time — Leontiskos

A.I. is significantly useful to me, because vetting its answers is not the same thing as knowing them beforehand. It can point me to an interpretation that I hadn’t thought of, and I can then verify the credibility of that interpretation.

We're asking a viewpoint question instead of a truth question, and that's part and parcel of the whole nature of AI. — Leontiskos

Isnt a philosophy a metaphysical viewpoint or worldview?

I merely emulate Wittgenstein, who rightly noted that a serious and good work of philosophy could be (and I would add has been) written consisting entirely of jokes. — Ciceronianus

That's Plato, one of the best philosophers ever. He's all jokes, all the way through, until you hit the "Laws", the most mundane and boring work ever, but that's more like dogma than philosophy.

Interacting with AI doesn't require other humans.
Anyone can do so on their own time; I mostly do by coincidence when doing google queries (I think).

Here at the forums, I kind of expect interacting with humans.
Or, at least, if interacting with AI by proxy / indirectly, that it's relevant (and checked) summaries or the like.

How to distinguish, though?

:D I could use a memory upgrade