Using AI in debates.

[an anarchist]

"Banned" feels a bit extreme (it's not as if anyone could stop you) but I think a better question would be why? What do you think the purpose of online debate is and how would just copying output from a chatbot support or achieve that?

You can't debate with a chatbot (AI or otherwise) because it isn't thinking about the topic in the same way. It's just compiling information from online sources that relate (or appear to) the topic. If you're going to do that, you might as well just link the original human-written information making those arguments directly.

Obviously I want your answers here, not the output from a chatbot. If I wanted to "debate" with a chatbot, I'd be on a chatbot site rather than a discussion forum.

I understand your concerns and I appreciate your point about the distinction between an AI and a human interlocutor.

In another online venue I was going to participate in, I found myself in a situation where I had crafted an opening post presenting my arguments with the assistance of AI, only to discover upon triple-checking the rules that the use of AI was prohibited. Although it may be difficult for sites and individuals to reliably detect when AI is being used in an online debate, I believe in respecting the rules of the platforms I participate in. That's why I think it's important to have this conversation here on RF before any rules are established regarding AI use.

As for the "why?" you posed, purposes for online debate can be multifaceted. For some, it's about refining their own viewpoints through rigorous discussion; for others, it's about persuading others to their point of view. AI can be a useful tool in achieving these goals by providing factual information, presenting a range of arguments and viewpoints, and helping to articulate thoughts more effectively.

To reassure you, it's very much me you're talking to. The "online version" of any of us is, to some extent, a version. We choose how we present ourselves, what we say, and how we say it. When I use AI to assist me in online debates, it's merely another layer in that process—an enhancement that allows me to communicate more effectively. I'm still the one making the decisions, setting the direction of the debate, and providing the insights. The AI is a tool, no different from a search engine or a reference book, that I use to help me achieve my goals in the debate.

 

[wellwisher]

There was a study at MIT about how AI and humans think differently. When the human brain writes to memory it adds emotional tags. Our memory has both sensory content and emotional tags. This allows us to use both sides of the brain to process our memory; left=sensory data and the right=emotions. AI will make use of the sensory content; cold hard facts, but since it does not use emotional tags, it is not as good at subtle spatial nuance that that emotional tagging process allows.

In the MIT study, they asked humans and AI the same questions about enforcing punishment for breaking the rules. The humans were more flexible, since they would not just use the obvious cause and affect with any arbitrary set of rules. Humans would also empathize with unique situations that may not show up with just the known cold hard facts. The AI is a colder task master.

If you look at the evolution of computers, much of what we do today in terms of interfacing with a computer, such as voice command, was originally designed for the blind; handicapped. Now we are all using the handicapped ways, due to the ease; less need for hand eye coordination that the blind lacked. AI could be appealing but it may also handicap you, if you start to get overly dependent on a machine doing things for you. I suppose the captured audience is a good marketing approach.

There is also what can be called the oracle of the false prophet. As AI weakens the mind, by thinking for you, some humans will eventually make it an oracle, with a particular trained machine or program being the oracle of a false prophet.

The analogy is buying some new cyborg leg covers, that allow you to walk and run faster. Since it is so much easier and faster to get around, you wear them all the time and begin to depend on them to get around. This causes your natural legs to atrophy and become less useful. This tipping point changes your fun loving desire, to a deep need for the false legs to get around. Now their importance is irrationally amplified beyond it original playful use. My advice is do not wear them all the time, but maintain natural also; have fun instead of creating a need.

 

[an anarchist]

I like it when people need extensions of their natural intelligence to debate me. That means they are losing.
Especially when they use, as help, what is nothing more but a very complex and sophisticated unthinking parrot, like chatGPT.

Ciao

- viole

The usage of AI as a tool in an online debate does not necessarily indicate a winning or losing stance. Instead, it's about maximizing the effectiveness of one's argument. To imply that the use of an AI assistant means someone is "losing" is akin to saying that a scholar using a reference book or a search engine to supplement their argument is on the losing side. These tools, whether traditional or advanced like AI, are used to enhance and inform our discussions, not as a signal of defeat.

Furthermore, intelligence is not binary or simple. The usage of AI does not automatically denote the intellectual capacity of the user. Just as using a calculator doesn't necessarily mean someone is bad at math, using AI doesn't imply that someone is less capable. Instead, it's an indication that they're utilizing the resources available to them to make their argument as robust and well-informed as possible.

As for your metaphor of AI being a "sophisticated unthinking parrot", while it's true that AI doesn't think or understand in the way humans do, it's important to remember that AI is not simply parroting information. It's analyzing vast amounts of data, identifying patterns, and producing responses based on that analysis. This is a far cry from simple repetition. It's a sophisticated process that can provide valuable insights and information, contributing to the quality of our debates.

 

[dybmh]

I like it when people need extensions of their natural intelligence to debate me. That means they are losing.
Especially when they use, as help, what is nothing more but a very complex and sophisticated unthinking parrot, like chatGPT.

Ciao

- viole

And if you cannot refute the parrot?

And any book, any resource, any website, is an extension to a person's natural intelligence. So, that's a silly argument.

 

[mikkel_the_dane]

And if you cannot refute the parrot?

And any book, any resource, any website, is an extension to a person's natural intelligence. So, that's a silly argument.

Even the parrot as the broken clock could be right.

 

[an anarchist]

Well, ask it what meta-cognition is for all its variants and how that relates to say information.
All of this is in effect meaningless unless we as we learn to be critical of even information as an idea.

(Personal, non AI generated note: We'll have to ask it again in a couple months. It can only browse the web for a limited amount of time in its present state. Of course, I can change the settings and just feed it direct online links now and perhaps even digital documents in the near future. So there could be a way for me to get a better answer from the AI just gotta change some settings. So such questions like the ones you tend to ask I'm sure will have a more satisfactory response than the one I am going to post right now.)

Metacognition is the practice of being aware of one's own thinking. It's often referred to as "thinking about thinking." When we notice ourselves having an inner dialogue about our thinking and it prompts us to evaluate our learning or problem-solving processes, we are experiencing metacognition at work. This skill helps us think better, make sound decisions, and solve problems more effectively. Metacognition comprises three key aspects: planning, monitoring, and evaluation.

1. **Planning Strategies:** This involves anticipating the strengths and weaknesses of ideas, scrutinizing plans at a time when they can most easily be changed.

2. **Monitoring Strategies:** These help students check their progress and review their thinking at various stages. It allows for adjustments while the plan, activity, or assignment is in motion.

3. **Evaluation Strategies:** Much like the mirror in a powder compact, evaluation strategies serve to magnify the image, allowing for careful scrutiny. When students inspect parts of their work, they learn about the nuances of their thinking processes.

As for how metacognition relates to information, metacognition is critical for processing and understanding information effectively. Metacognitive strategies, such as self-monitoring and reflection, allow individuals to better manage and interpret information. They provide the means to evaluate the reliability and relevance of information, make connections between different pieces of information, and apply information to different contexts. By reflecting on how they are understanding and using information, individuals can improve their learning processes and outcomes.

As for your statement about being critical of information as an idea, I agree that it is essential. Our understanding and use of information should not be passive. Instead, we should critically evaluate the sources of our information, the methods by which it was obtained, and the assumptions underlying it.

I encountered some difficulties while attempting to provide a more detailed explanation of the concept of "information." Thus, I'm unable to elaborate on that at this time. However, generally, "information" can be understood as data that is processed, organized, structured, or presented in a given context so as to make it useful or meaningful. It is what is conveyed or represented by a particular arrangement or sequence of things.

Your comment about the need to be critical of even information as an idea is very insightful. It's indeed crucial to question and scrutinize the nature and sources of the information we receive, the methods by which it was gathered, and the contexts in which it is presented or used.

 

[Father Heathen]

I would love to see someone ask the ai for sources and then accuse it of being biased and one sided.

Given that A.I. uses logic, it will undoubtedly be accused of liberal bias.

 

[dybmh]

Yes. If a claim is predicated on the validity of classical logic, and it is derived using classical logic, then the claim is tautologically true, within classical logic. By definition.

Doubting this, would be the intellectual equivalent of doubting the marital status of bachelors.

Ciao

- viole

And what tools does one use to predicate that validity? It cannot be classical logic, because classical logic does not have the tools to evaluate evidence nor relelvance.

 

[mikkel_the_dane]

(Personal, non AI generated note: We'll have to ask it again in a couple months. It can only browse the web for a limited amount of time in its present state. Of course, I can change the settings and just feed it direct online links now and perhaps even digital documents in the near future. So there could be a way for me to get a better answer from the AI just gotta change some settings. So such questions like the ones you tend to ask I'm sure will have a more satisfactory response than the one I am going to post right now.)

Metacognition is the practice of being aware of one's own thinking. It's often referred to as "thinking about thinking." When we notice ourselves having an inner dialogue about our thinking and it prompts us to evaluate our learning or problem-solving processes, we are experiencing metacognition at work. This skill helps us think better, make sound decisions, and solve problems more effectively. Metacognition comprises three key aspects: planning, monitoring, and evaluation.

1. **Planning Strategies:** This involves anticipating the strengths and weaknesses of ideas, scrutinizing plans at a time when they can most easily be changed.

2. **Monitoring Strategies:** These help students check their progress and review their thinking at various stages. It allows for adjustments while the plan, activity, or assignment is in motion.

3. **Evaluation Strategies:** Much like the mirror in a powder compact, evaluation strategies serve to magnify the image, allowing for careful scrutiny. When students inspect parts of their work, they learn about the nuances of their thinking processes.

...

But that is even limited in the sense that is about learning. My wife uses it as a work tool in her line of work.
And I use it when I do skepticism.

 

[Yerda]

I appreciate your perspective, but it appears we approach this topic from fundamentally different mindsets.

Let me share an anecdote: in school, I would sometimes use a chessbot to play against my friends without their knowledge. My goal was to win and enjoy the benefits that come with that victory. This wasn't because I didn't appreciate a genuine game of chess—quite the opposite, I enjoy a challenging match. But there's a certain satisfaction in achieving a swift victory, even if it's attained through unconventional means.

Translating this to the realm of online debates, the goal isn't always just to engage in intellectual sparring for its own sake. Sometimes, the aim is to disseminate and propagate crucial information effectively, which can be facilitated by "winning" a debate.

In saying this, I'm not suggesting that everyone should adopt my approach, or that my way is inherently superior. We all have our own reasons for participating in these debates. Personally, I engage in debates to learn, and when I believe I'm right, I debate to win.

Using AI in this context doesn't detract from these goals—in fact, it can enhance them by providing stronger, more well-rounded arguments and by ensuring that the information I present is accurate and comprehensive.

So while I understand your perspective, I hope this gives you some insight into why some of us might see the value in incorporating AI into our online debates.

If a chatbot helps you to enjoy the forum, by all means, go for it.

 

[dybmh]

Even the parrot as the broken clock could be right.

The problem that I'm encountering with AI, is that it is forgetful. I think this is a built in feature/limitation. I've asked it what these limitations are, and it doesn't know. So any lengthy debate with an AI would become erroneous quickly.

But it is excellent at looking for flaws in something I am saying. As a method of error-checking it's very good. But it only presents a list of possible weaknesses. Then each of these can be analyzed individually.

Of course, it doesn't know that those weaknesses are sometimes defeatable. And because it is forgetful, it loses the necessary conditions. But with practice a person can formulate questions that deliver the desired error checking.

I've been using the phrase "without revision" frequently. Also, "necessary condition". Example: Please hold the following necessary conditions: 1,2,3 and evaluate the following proof for errors and weaknesses without revision? Then I ask if the proof is rigorous and sound. Then I ask if there are any possible weaknesses no matter how small. Then I ask it to write a proof to defeat it.

I revise and retest repeatedly so that these questions can be asked and answered quickly before the AI has forgotten anything. That's been working really well.

 

[an anarchist]

The problem that I'm encountering with AI, is that it is forgetful. I think this is a built in feature/limitation. I've asked it what these limitations are, and it doesn't know. So any lengthy debate with an AI would become erroneous quickly.

But it is excellent at looking for flaws in something I am saying. As a method of error-checking it's very good. But it only presents a list of possible weaknesses. Then each of these can be analyzed individually.

Of course, it doesn't know that those weaknesses are sometimes defeatable. And because it is forgetful, it loses the necessary conditions. But with practice a person can formulate questions that deliver the desired error checking.

I've been using the phrase "without revision" frequently. Also, "necessary condition". Example: Please hold the following necessary conditions: 1,2,3 and evaluate the following proof for errors and weaknesses without revision? Then I ask if the proof is rigorous and sound. Then I ask if there are any possible weaknesses no matter how small. Then I ask it to write a proof to defeat it.

I revise and retest repeatedly so that these questions can be asked and answered quickly before the AI has forgotten anything. That's been working really well.

AI models like GPT-3 and GPT-4 indeed have a "context window", which is the maximum number of tokens (individual pieces of information, such as words or punctuation) that the model can consider at once when generating a response. This limitation is due to the design of the underlying transformer architecture which these models use. If a conversation exceeds this limit, the model won't be able to see or remember the earlier parts of the conversation.

For GPT-3, the context window is 2048 tokens, while for the standard GPT-4, the context window is significantly larger, at 8000 tokens. There's also an extended version of GPT-4 with a context window of 32,000 tokens. This extended context window can be particularly beneficial in maintaining longer and more intricate conversations, as it lessens the chance of the AI "forgetting" earlier parts of the conversation due to exceeding the context window limit.

That said, as your current conversation partner, I can remember the entire conversation as long as the earliest message in this current conversation that I can see is your request for me to construct an opening post. The AI's ability to maintain context over extended conversations is one of the reasons why many users find value in using services like GPT-4, as it allows for deeper and more nuanced interactions.

 

[9-10ths_Penguin]

I wanted to make a thread debating whether AI should be banned in online debate.

I am going to use AI starting after this sentence for the rest of this thread to argue in favor of allowing AI to be used in online debate.

Subject: The Role of AI in Enhancing Online Debate - A Case for AI Integration

Hello fellow debaters,

The role of artificial intelligence (AI) in various aspects of life is growing significantly. As we find ourselves in the digital age, the debate surrounding the use of AI in online discussions is timely and essential. It is my belief that AI should be allowed to enhance our online posts in a debate setting, and here's why.

Firstly, the integration of AI in online debates promotes an informed and educated discourse. AI, like OpenAI's GPT-3, can provide quick access to an enormous amount of information. This ability can be leveraged by users to fact-check their arguments, provide relevant statistical data, or even generate well-structured arguments based on the given inputs. This not only helps in making the discussions more fact-based but also contributes to the overall quality of the debate.

Secondly, AI can help in reducing the cognitive load of participating in online debates. Crafting thoughtful and articulate responses requires time, effort, and knowledge. AI can assist users in formulating their thoughts, enabling them to participate in multiple debates simultaneously without compromising the quality of their contributions.

Moreover, AI could level the playing field for those who may not be as confident or articulate in their writing. By providing linguistic and rhetorical assistance, AI can help less experienced users to express their ideas clearly and persuasively. This democratization of debate skills could lead to a richer diversity of voices being heard.

Critics may argue that AI use might lead to inauthentic debates or could potentially be manipulated to spread misinformation. However, these are issues of regulation, not inherent flaws of AI. As with any tool, appropriate guidelines and oversight can ensure its ethical use. And it's also worth noting that AI systems can be designed to cite their sources, increasing transparency and allowing users to scrutinize the information being presented.

In conclusion, AI has the potential to greatly enhance the online debating experience. By promoting informed discourse, reducing cognitive load, and democratizing debate skills, AI could usher in a new era of online discussion. However, to realize these benefits, we must be open to its integration and proactive in establishing rules for its use.

Looking forward to hearing your thoughts and counterarguments on this fascinating subject.

Best,
@an anarchist

If someone isn't into online debates, why would they participate at all?

If they are into online debates, why would they outsource what they do - ostensibly for fun - to an AI?

 

[9-10ths_Penguin]

What kind of debates would AI have? If there's a local debate over whether to build a new bridge at Elm St. or Main St., would AI be able to take a side in such a debate?

If you ask it to take a side, it will take one... and pull from all sorts of online sources - reliable or not, verified or not - to back up its position.

 

[PureX]

I like that AI can quickly list and pose the alternative views on a debate subject. That can be useful to the human debate participant, certainly. But beyond that, it’s a pointless exercise. The machine doesn’t get any smarter. It just gets better at oppositional response.

 

[9-10ths_Penguin]

In another online venue I was going to participate in, I found myself in a situation where I had crafted an opening post presenting my arguments with the assistance of AI, only to discover upon triple-checking the rules that the use of AI was prohibited. Although it may be difficult for sites and individuals to reliably detect when AI is being used in an online debate, I believe in respecting the rules of the platforms I participate in. That's why I think it's important to have this conversation here on RF before any rules are established regarding AI use.

Any time someone presents material they didn't write as their own, it's plagiarism. Even if it doesn't violate some official forum rule or the like, it's dishonest. AI doesn't change this.

If you want to quote what ChatGPT has to say about a subject - and clearly acknowledge that ChatGPT generated it - then go nuts... thought this will probably be about as compelling as saying "here's what some random blogger who I'm not sure is knowledgeable wrote about the topic."

If you're only using the AI to generate a reading list on a subject, then you go through the sources it provides, check each one, and use your notes to create your original post, then I don't see a problem with this.

 

[HonestJoe]

AI can be a useful tool in achieving these goals by providing factual information, presenting a range of arguments and viewpoints, and helping to articulate thoughts more effectively.

Not in the way you're suggesting I think. You can't guarantee that any information produced by AI chatbots is factual or any arguments are fact based, logical and internally consistent. It's bad enough with all the humans posting false information and flawed arguments (intentionally or not) without adding a whole load more artificially.

The AI is a tool, no different from a search engine or a reference book, that I use to help me achieve my goals in the debate.

AI is not a tool, it is a technology which can be used to develop a whole range of tools (a bit like the difference between an engine and all the different vehicles and machines an engine can be part of).

You're free to use any kind of AI tools (or indeed non-AI tools) to help develop your opinions and arguments on the forum, but that shouldn't make any difference to the overall manner the debate happens. You still need to understand the arguments you make and if you make claims or assertions, you need to be able to support them. "My AI told me" isn't going to cut it, even if it did.

 

[syo]

I wanted to make a thread debating whether AI should be banned in online debate.

I am going to use AI starting after this sentence for the rest of this thread to argue in favor of allowing AI to be used in online debate.

Subject: The Role of AI in Enhancing Online Debate - A Case for AI Integration

Hello fellow debaters,

The role of artificial intelligence (AI) in various aspects of life is growing significantly. As we find ourselves in the digital age, the debate surrounding the use of AI in online discussions is timely and essential. It is my belief that AI should be allowed to enhance our online posts in a debate setting, and here's why.

Firstly, the integration of AI in online debates promotes an informed and educated discourse. AI, like OpenAI's GPT-3, can provide quick access to an enormous amount of information. This ability can be leveraged by users to fact-check their arguments, provide relevant statistical data, or even generate well-structured arguments based on the given inputs. This not only helps in making the discussions more fact-based but also contributes to the overall quality of the debate.

Secondly, AI can help in reducing the cognitive load of participating in online debates. Crafting thoughtful and articulate responses requires time, effort, and knowledge. AI can assist users in formulating their thoughts, enabling them to participate in multiple debates simultaneously without compromising the quality of their contributions.

Moreover, AI could level the playing field for those who may not be as confident or articulate in their writing. By providing linguistic and rhetorical assistance, AI can help less experienced users to express their ideas clearly and persuasively. This democratization of debate skills could lead to a richer diversity of voices being heard.

Critics may argue that AI use might lead to inauthentic debates or could potentially be manipulated to spread misinformation. However, these are issues of regulation, not inherent flaws of AI. As with any tool, appropriate guidelines and oversight can ensure its ethical use. And it's also worth noting that AI systems can be designed to cite their sources, increasing transparency and allowing users to scrutinize the information being presented.

In conclusion, AI has the potential to greatly enhance the online debating experience. By promoting informed discourse, reducing cognitive load, and democratizing debate skills, AI could usher in a new era of online discussion. However, to realize these benefits, we must be open to its integration and proactive in establishing rules for its use.

Looking forward to hearing your thoughts and counterarguments on this fascinating subject.

Best,
@an anarchist

People like long essays, eh? They bore me.

 

[Twilight Hue]

People like long essays, eh? They bore me.

So much so they pull the plug...

Facebook Shuts Down AI Robot After It Creates Its Own Language

When English wasn't efficient enough, the robots took matters into their own hands.

www.huffpost.com

 

[an anarchist]

If someone isn't into online debates, why would they participate at all?

If they are into online debates, why would they outsource what they do - ostensibly for fun - to an AI?

I understand your perspective, but I'd like to suggest a different viewpoint. Using AI in online debates isn't necessarily "outsourcing". Instead, I see it as a tool that enhances my own thoughts and arguments. The AI doesn't replace my role in the debate; rather, it aids me by offering different perspectives, helping to identify potential flaws or strengths in arguments, and assisting in articulating my thoughts more coherently.

Moreover, there's a certain enjoyment to be derived from employing AI in this way. It's like discovering a new tool or technology that offers new ways of thinking and communicating. In fact, I've found it quite fun and engaging to use AI in this context, and it's an avenue I've only just begun to explore. I believe the use of AI can add a unique layer to the process of online debating, enhancing rather than diminishing the personal involvement and enjoyment inherent in these discussions.