I meant to question this:
I’m not sure this is right.
In what way, do your ideas change?
If you notice that AI has not expressed your ideas correctly, then you point it out and it will be sorted!
If it happens to produce an idea that is useful, then you accept and acknowledge?
Or is that too complicated?
Re: ‘expression’, I think that needs further explanation.
I feel like these horses were beaten to death on the previous forum, but:
There is an equivocation on “argument” in this statement, and AI enthusiasts are constantly involved in these sorts of equivocations. AI does not make arguments; it is just a fancy repeater-machine.
Similarly, someone might convince themselves that they are having a “conversation” with a chatbot, but in time they come to realize that it is not a conversation at all. In time they realize that they are projecting words into a cultural-linguistic mirror. It’s much like convincing yourself that the mirror in your bathroom is a photographer, or that the reflection you see is itself a person. After all, from the right angle it does look like a person.
At the end of the day I would say that those who do not have the ability or the time to formulate their own ideas do not ultimately want to participate in a philosophy forum. They are using AI as a shortcut unto their real end: the propagation of a particular set of ideas. “I don’t have the ability/time to write these posts myself, so I will lean on a chatbot. After all, what really matters is that the ideas I value get widely propagated.” This is a species of what TPF pejoratively calls “evangelization.”
I think this is right. Isn’t there something already in the Guidelines re evangelical-type posters? Repetitive patterns like an obsession on a particular topic. The need to win the argument because it has become a major part of their identity.
Probably. But then, it seems never to be adequately resolved…
The content and interpretations of Wittgenstein’s ideas change depending on how they’re expressed. Scholars can therefore debate stuff like punctuation, comma-signs, choice of words, the translation from German etc
It is often the case that when I begin to write I do not know what am going to say. It is not a matter of putting my thoughts into words, but of thinking through writing.
I am of the opinion that if something cannot be expressed clearly it has not been thought clearly. If clear thinking is one’s goal then clear expression should be a means to that goal.
Yes. I know that ideas change, as do some philosophers at different times in their lives (early and late). I recognise the impact of experience, knowledge, debates and discussion. Relating and connecting.
I was referring to ChatGPT when asked to edit your text (written thoughts) — How can it change your ideas?
Perhaps, it does. Depending on the user. Unlikely if the user aims to persuade people of a static or dogmatic set of beliefs.
How much control does ChatGPT have? It is a tool, working with your permission.
You should write a book! 
I have sometimes had the experience of trying to express some idea and, realizing that the first version is off, changing the expression, and then realizing that’s off too, eventually concluded that I’m having trouble finding just the right words because what I’m trying to say is not true.
An LLM does not experience that friction, but can express any idea equally smoothly. By using an LLM, you deprive yourself of that valuable friction.
Copy/pasting AI outputs without it clearly marked as an output and for a good reason is a slippery slope that will eventually fill this forum with a lot of text from people lazy enough not to write themselves.
We don’t want to read through pages of AI generative meandering, I can ask an LLM myself if I want such stuff. If people can’t write their own stuff, they should go somewhere else.
One assumes it’s possible to paraphrase. For instance, I can tell you what the Cyclic Argument is without delivering it to you in the original Greek. It’s Plato’s ideas. The fact that it’s expressed in an utterance 2400 years later doesn’t change that.
It may be that from a certain point of view an utterance and its content are a unity. I personally wouldn’t use that as a basis for censoring you though. If I censor you it should be for a reason you understand. Otherwise I’m just a fool on a hill, chopping bytes for no reason that anyone will ever care about or remember.
Another example from philosophy is how Kant’s idea of ‘analyticity’ is changed by the way Carnap expresses it. The idea was attacked by Quine in his famous “Two Dogmas…”, but he attacks Carnap’s version, not Kant’s.
An AI does what Carnap does, changes the idea with its ways of expression.
Suppose someone is a tennis player. He sees that he can buy a robotic arm which swings itself and hits excellent shots. He buys the arm under the impression that he is improving the quality of the game. He then buys robotic legs that automatically move him around the court more effectively than he could do himself. He tells himself that he is improving the quality of tennis. He also begins to win.
The quality of the matches initially drops because of the significant mismatch between the robotic player and the others. But then, seeing this, the other players also invest in the robotic technology. The quality of matches now seems to increase, and crowds flock to the tennis stadium to watch. The ball is moving faster and the players are moving faster, and that’s the point of tennis, isn’t it?
Soon enough the human players are replaced by bona fide robots. Then, as technology exponentially increases, a way is devised to compact the raison d’être of tennis into the ball itself. A highly sophisticated ball now propels itself across the net with an unbelievable array of spins, angles, and speeds. The progress of tennis is now complete. The crowds at Wimbledon begin to dwindle. Fans scratch their heads, “Is this what we wanted? It does seem to be what we were aiming at each step of the way, but now that tennis is of the highest ‘quality’ it turns out we are not fans of tennis after all.”
I don’t disagree with TPF’s claim that human expression is a sine qua non of philosophy, and that “argument” (i.e. material symbols mimicking inference) is insufficient. But I would also say that, in a more general sense, the error of the AIers is to reduce philosophy to a single variable, which in their mind is called “quality of argument.” Their claim is this: “The only difference in my AI-generated post is improved argument quality, and quality of argument is the whole point of philosophy.”
Both parts of that claim are wrong. As to the first half, the nature of the post has changed in ways other than “quality,” just as the tennis player’s decision to use a robotic arm has many effects that he wishes not to see. As to the second half, philosophy is not reducible to a unitary criterion of “argument quality” (whatever that means).
If someone understood why a post should not be published, then there would be no need for censorship. If everyone understood and agreed with the reasons for the rules, then there would be no need for rules. Rules and censorship presuppose a lack of understanding in those to whom they are applied.
Interesting. It seems I have a lot to learn.
However, my understanding is that ChatGPT doesn’t change your ideas unless you ask it to.
ChatGPT always tries to help directly. It does not naturally consider alternative vantage points unless you ask for them. Changing the narrator or point of view in the prompt opens up new ways for the AI to consider answers, whether it’s through the eyes of an inanimate object, a pet, or simply from a long way away. ChatGPT responds by reorganizing information around the new narrator rather than rewriting the same answer with different adjectives. The effect is often surprising because it reveals angles you would not have considered on your own. ChatGPT has a hidden set of creativity switches — here’s how to use them | TechRadar
How did we get here? Leaving this, now, thanks.
I haven’t used ai generated text, just never got around to it, but I have been using AI music production.
Whether we would say a certain song is by the human user or by the AI depends on how the AI is used. AI music is riding on digital production technology that’s been around for decades (same with AI visual art). The difference is that these tools can now be used with a higher level interface.
One artist who is now gaining fame (Teddy Swims) admits that he uses AI to listen to his songs in different genres. It’s his voice, and his melody, but AI can give him different versions: country, hip hop, old rock and roll, etc. After using Ai, I understand why all of his songs are very tight, and in spite of ranging over the map of music styles, he nails every genre to a t. It’s not that he couldn’t have done this without AI, it’s just that AI makes it easily available to anyone who can pay to use the software. At first, I was put off, but now I’m finding a way into appreciating it for what it is.
So I wonder if the same is true of AI text. Maybe some usage amounts to spell checking and flow recommendations that have been around forever. I can’t think of a good reason to reject that kind of use.
On the other hand, if you leave the prompt space blank in an AI music generator, it will blow your hair back with a complete song with lyrics and singer. This is why they’re being sued. If someone uses AI to completely construct a reply and passes it off as their own, that’s problematic.
If you ask AI to summarize a large body of work, then obviously the product is going to be AI generated material.
A good philosopher will have a large body of high-quality content. And study of that content will reveal how the author’s ideas evolved over time. That’s why Plato for example is commonly analyzed as three phases. What would be the purpose of the author reproducing the same material over and over again?
That’s exactly the case, but it’s not a difficulty, c’est très facile. The rule is about authorship. It’s not a rule about quality, there’s other rules for that.
The key element is the experience of not being satisfied with certain expressions one makes. The composition of speech and writing involves putting forward statements and striving to either refine or correct them. The pretense of AI does not bear upon the creation of ideas but the relationship of thinking to suffering.
I’m reluctant to use AI for creative work, because it’s not just like any tool or software with which I can produce, improve, or remix my stuff. I type prompts and select the results that I wish to keep. But the results are more or less composed of other people’s stuff on which the AI has been “trained” or plagearises. In this sense I’m neither the author nor composer, but a plageariser getting away with it. Also my prompts and selections “train” the AI, so whatever value or competence there is in my use, it can now be used by the AI provider. Currently I’m using AI mostly as a search engine or encyclopedia.
I’d say a person hasn’t begun to realize the potential of AI visual art until they start supplying their own images and putting in clashing prompts. It’s not plagiarism. It’s something entirely new.
Same thing with music. If you don’t put in your own melodies, you’ll tend to get elevator music.