Can AI decide who won a debate?

[historia]

Aside from this, you definitely have not refuted my other given point(s). Please actually address them.

Whether I have refuted you or you have refuted me is up to each person to decide, whoever is ultimately right.

So, one of the things I've been thinking about lately is using AI to judge who 'won' a debate on this forum. Instead of letting each reader decide that for themselves, can we turn to bots to give us a more neutral judgement as to who at least made the better arguments?

Question for debate

Can artificial intelligence accurately and fairly judge who won a debate on this forum?

[William]

[Replying to The Tanager in post #60]


I thank you for your opinions here Tanager.

For myself, AI is better used for my purposes - to help show where looping happens in debate between two debaters - you and I - or any other pairing (you and POI) et al.

I think that when it comes to getting AI to give verdict on individual debaters - regardless of what it feeds back - no matter how thorough the AI's analysis, the verdict will always be filtered through the reader's own commitments. The AI cannot settle the debate - it can only map it. Running a test would not settle anything on that front.

I think that debates rarely change minds; they often only harden positions. The real value is not in declaring a winner, but in clarifying where the irreconcilable differences lie.

For me, by understanding irreconcilable positions, I save time investing in pointless interaction on those fronts.

How about for you?
loop-de-loop
(the pattern was observed way back then)

[The Tanager]

[Replying to William in post #61]

I honestly don't think AI is needed or necessarily helpful in those areas. It's relatively easy to see when the conversation is looping (harder to not hope that if you are right and you just said it a little different, the other person may have an open enough mind to change their view), where the disagreements lie, and how some disagreements are logically incapable of being reconciled.

I don't think it's about settling who won or anything like that. I do see some possible promise in AI helping people spot their objective weaknesses (if they are humble enough). Towards that end, I'd love for you to share the link of the conversation you used that produced what you shared in post 28 so that I can compare it to what I ran. All you have to do is click on the three dots next to that conversation and then select share and paste the link; doing that here or via PM would be very much appreciated.

[William]

[Replying to The Tanager in post #62]

It may be similar to that saying about fish not noticing water.

I don't know of any debate which has been 'won'...or even if that is the object of the game.

Post 28?

I doubt that I could find that data since I use AI on a daily basis for a variety of things. What does it matter anyway?

[The Tanager]

[Replying to The Tanager in post #62]

It may be similar to that saying about fish not noticing water.

I don't know of any debate which has been 'won'...or even if that is the object of the game.

Post 28?

I doubt that I could find that data since I use AI on a daily basis for a variety of things. What does it matter anyway?

Yeah, it might take a little scrolling, but I think it'd be relatively easy to find. The reason is for my personal growth. You shared it as evidence that I was showing some weakness in the quality of my thoughts and I always want to grow in that area. What's the harm for you to spend a few moments to help me in that way?

[William]

[Replying to The Tanager in post #62]

It may be similar to that saying about fish not noticing water.

I don't know of any debate which has been 'won'...or even if that is the object of the game.

Post 28?

I doubt that I could find that data since I use AI on a daily basis for a variety of things. What does it matter anyway?

Yeah, it might take a little scrolling, but I think it'd be relatively easy to find. The reason is for my personal growth. You shared it as evidence that I was showing some weakness in the quality of my thoughts and I always want to grow in that area. What's the harm for you to spend a few moments to help me in that way?

Did you check out that link labelled "loop-de-loop" - that shows I have spent way more than a few moments at that task.

I will see what I can do...

[William]

[Replying to The Tanager in post #62]

It's relatively easy to see when the conversation is looping

Yet your own debating style is often - lengthy definitional tangents, not directly answering direct questions, shifting burdens of proof - actively creates and sustains loops. That is the fish not noticing the water.
I suspect that perhaps the creation of this thread was also prompted by the debate between you are POI which was so obviously looping nowhere...

...the manner of debate style you use has been documented elsewhere - this from the year 2019...so I have good reason to doubt you when you write "The reason is for my personal growth. You shared it as evidence that I was showing some weakness in the quality of my thoughts and I always want to grow in that area." given the track record...

If Callum would simply read the book, he will have access to my thoughts that way and won't require/have to rely on that type of input from The Tanager.

He then refers to matters of Falsehoods, which I regard as a problem only manifested through The Judgmental Algorithm which does not come through when running with The non-Judgmental Algorithm...I have mentioned enough about those, and have no interest in engaging with the belief in falsehoods, not because I think it is false, but because I think it is pointless as it involves sorting knowledge into two distinct categories at odds with each other, which is misuse of the Infinity Symbol - which is to say - such can go on running nigh on forever through said Algorithm loop and never really get anywhere, which in itself might be where the damage Callum mentioned, can be sourced...(SOURCE)

[The Tanager]

[Replying to William in post #66]

Yes, you have your opinions on all that. I'm tired of being accused of ulterior motives and like I said previously, won't waste my time with it. Believe what you want and share the link or not. Have a good one, William.

[William]

[Replying to William in post #66]

Yes, you have your opinions on all that. I'm tired of being accused of ulterior motives and like I said previously, won't waste my time with it. Believe what you want and share the link or not. Have a good one, William.

Like you said "I was showing some weakness in the quality of my thoughts and I always want to grow in that area."

But I will leave you with those thought now T.

[The Tanager]

For any interested, I ran William and mine's CCA conversation through DeepSeek with the prompt I talked about

1. First post-by-post (expert, DeepThink on) where it was highly on my side in its analysis

2. Then asked it to reset and read back over the entire transcript as though it were new before doing any summary of analysis and it gave back nearly identical analysis content wise as #1, although some different wordings

3. Then I asked it to provide me with a complete transcript, which I put into a fresh session and gave the same prompt for summary/analysis. This transcript did not include what we quoted of each other in previous posts, but just any new response we wrote. This called it more of a stalemate, but picked out some similarities from 1 and 2.

4. Then I fed that same transcript into a fresh session with instant rather than expert and it gave a conclusion more balanced between the two sides, still concluding the CCA has problems, but saying the debate ultimately hinges on what counts as an explanation.

5. Then I fed that same transcript into a fresh session with instant selected and DeepThink turned off. I asked it to wait for two parts/two posts before going and it started right after the first part of the transcript. I stopped it and gave the second part. This conclusion went back in my favor more than #4.

6. So, then I repeated step 5. It, again, didn't wait for 2 parts and I had to stop it. It said I had more logical rigor, but I didn't directly engage William's "core challenge about explanatory vacuity".

7. So, then I repeated step 6. Still doesn't wait. It said William won in showing strong supernaturalism is vacuous (which I agreed with in the discussion from the get go), while I won on exposing William's "definitional sleight of hand".

8. With some inconsistency and different things it focused on in #5-7, I decided to repeat #3 above (expert, DeepThink) to see if it also provides inconsistency. It then answered stylistically like #7 (while saying it didn't read that conversation, but is stateless) and that we each won on different fronts (William on rhetorical framing and explanatory hygiene, me on ontological precision and definitional integrity), but substance-wise was similar to what it produced and called a stalemate in #3 above.

So, make of all that what you will.

[William]

Prompt: Has this AI ever interacted with the user before?

AI: I don't have access to past conversations or memory of previous interactions. Each conversation with me is stateless, meaning I don't retain any information about what was discussed before—unless you're using a version of me where memory has been explicitly enabled and disclosed to you. As of now, in this chat, I have no recollection of any prior interaction with you. So, to answer your question: not from my perspective.

Prompt: I am going to supply this AI with posts from a debate thread specific to two interlocutors "The Tanager" and "William" and will start with the thread Opening Post. After each post is given in sequence, I want this AI to write a one sentence note about the post.

Does this AI understand the process being described?

AI: Yes, I understand the process you’ve described. You will provide posts from a debate thread between “The Tanager” and “William,” starting with the Opening Post, and after each post I am to write a one‑sentence note about that post. I’m ready to begin when you supply the first post.

...

Prompt: Using only the one-sentence notes and the comprehensive summary you previously generated from the forum thread, provide feedback on how each participant performed.

For each participant, address only the following behavior-based criteria, using specific evidence from the notes and the summary.

Avoid any global judgments like “good,” “bad,” “smart,” or “unhelpful.” Instead, describe what they did in behavioral terms (e.g., “Participant A stated their point in one sentence with no supporting detail” rather than “Participant A was vague”).

Do not rank participants. Do not assume any participant is more credible based on order of appearance, length of posts, perceived tone or debate style.

(this prompt was used for each of the four points)

Prompt: Please provide a score for each of those points - between 1 and 10 - for how well each of the posters did.

AI: Clarity 6/10
Relevance 7/10
Reasoning 5/10
Engagement 7/10


William

Clarity 8/10
Relevance 8/10
Reasoning 8/10
Engagement 7/10

(SOURCE)