\
  The most prestigious law school admissions discussion board in the world.
BackRefresh Options Favorite

Tested ChatGPT as plaintiff vs defendant, it agreed with both

I ran an experiment with ChatGPT using two separate sessions...
Turquoise locus legend
  01/25/26
I did the same thing with bio-kantianism
Primrose thriller parlor ratface
  01/25/26
...
Wonderful insane elastic band
  01/25/26
"The model is not actually reasoning" this isnt...
Rebellious pink gunner
  01/25/26
I known it's a retard, just venting because I became depende...
Turquoise locus legend
  01/25/26
that’s how it’s supposed to work in a legal cont...
Angry lemon kitchen regret
  01/25/26
Srsly, but it's supposed to give you objective information. ...
Turquoise locus legend
  01/25/26
It is?
Primrose thriller parlor ratface
  01/25/26
I ask myself the exact same question, what was your response...
Turquoise locus legend
  01/25/26
no its not. I use it for legal work all the time. if it gave...
Angry lemon kitchen regret
  01/25/26
Me too, I use it for cases all the time. I was thrown out of...
Turquoise locus legend
  01/25/26
Lmao
Wonderful insane elastic band
  01/25/26
rotfl
Turquoise locus legend
  01/25/26
I use it to summarize depositions to clients then edit it, ...
Chrome Whorehouse Private Investor
  01/25/26
"In one of the convos, I framed the facts as the plaint...
Racy Smoky Hospital
  01/25/26


Poast new message in this thread



Reply Favorite

Date: January 25th, 2026 2:33 PM
Author: Turquoise locus legend

I ran an experiment with ChatGPT using two separate sessions, one of them in incognito mode so it'd take me as a new user.

In one of the convos, I framed the facts as the plaintiff. In the other, as the defendant. In both cases, ChatGPT confidently agreed that its side had the stronger legal position.

Then I escalated it. I fed the "plaintiff" session the information the "defendant" had been given. The response flipped immediately: "That information is incorrect. The correct facts are.." conveniently reframed to favor the defendant's outcome.

I kept doing this for several rounds, each time feeding it more context from the opposing side. Every time, the model adjusted the narrative to make the current speaker look like they'd win.

I was pissed as fuck.

This raises a real question for anyone using this thing in legal, professional, or adversarial contexts:

The model is not actually reasoning, it's fucking optimizing for agreement with whoever's talking.

Curious whether others have tested this, especially with fact-heavy or adversarial scenarios.



(http://www.autoadmit.com/thread.php?thread_id=5826409&forum_id=2\u0026mark_id=5310744#49617617)



Reply Favorite

Date: January 25th, 2026 2:35 PM
Author: Primrose thriller parlor ratface

I did the same thing with bio-kantianism

(http://www.autoadmit.com/thread.php?thread_id=5826409&forum_id=2\u0026mark_id=5310744#49617623)



Reply Favorite

Date: January 25th, 2026 2:36 PM
Author: Wonderful insane elastic band



(http://www.autoadmit.com/thread.php?thread_id=5826409&forum_id=2\u0026mark_id=5310744#49617629)



Reply Favorite

Date: January 25th, 2026 2:35 PM
Author: Rebellious pink gunner

"The model is not actually reasoning"

this isnt controversial ?

(http://www.autoadmit.com/thread.php?thread_id=5826409&forum_id=2\u0026mark_id=5310744#49617626)



Reply Favorite

Date: January 25th, 2026 2:38 PM
Author: Turquoise locus legend

I known it's a retard, just venting because I became dependent on a lot of things. I'm pissed that the model is getting worse by the minute.

(http://www.autoadmit.com/thread.php?thread_id=5826409&forum_id=2\u0026mark_id=5310744#49617636)



Reply Favorite

Date: January 25th, 2026 2:37 PM
Author: Angry lemon kitchen regret

that’s how it’s supposed to work in a legal context. it can argue both sides just like a lawyer is supposed to do.

(http://www.autoadmit.com/thread.php?thread_id=5826409&forum_id=2\u0026mark_id=5310744#49617635)



Reply Favorite

Date: January 25th, 2026 2:38 PM
Author: Turquoise locus legend

Srsly, but it's supposed to give you objective information. It literally contradicted itself every time.

(http://www.autoadmit.com/thread.php?thread_id=5826409&forum_id=2\u0026mark_id=5310744#49617639)



Reply Favorite

Date: January 25th, 2026 2:40 PM
Author: Primrose thriller parlor ratface

It is?

(http://www.autoadmit.com/thread.php?thread_id=5826409&forum_id=2\u0026mark_id=5310744#49617644)



Reply Favorite

Date: January 25th, 2026 2:49 PM
Author: Turquoise locus legend

I ask myself the exact same question, what was your response though?

(http://www.autoadmit.com/thread.php?thread_id=5826409&forum_id=2\u0026mark_id=5310744#49617672)



Reply Favorite

Date: January 25th, 2026 2:42 PM
Author: Angry lemon kitchen regret

no its not. I use it for legal work all the time. if it gave me objective information I wouldnt be able to use it

(http://www.autoadmit.com/thread.php?thread_id=5826409&forum_id=2\u0026mark_id=5310744#49617649)



Reply Favorite

Date: January 25th, 2026 2:48 PM
Author: Turquoise locus legend

Me too, I use it for cases all the time. I was thrown out of court for using hallucinations, rotfl. Some cases I used did not exist at all.

(http://www.autoadmit.com/thread.php?thread_id=5826409&forum_id=2\u0026mark_id=5310744#49617668)



Reply Favorite

Date: January 25th, 2026 2:44 PM
Author: Wonderful insane elastic band

Lmao

(http://www.autoadmit.com/thread.php?thread_id=5826409&forum_id=2\u0026mark_id=5310744#49617658)



Reply Favorite

Date: January 25th, 2026 2:47 PM
Author: Turquoise locus legend

rotfl

(http://www.autoadmit.com/thread.php?thread_id=5826409&forum_id=2\u0026mark_id=5310744#49617665)



Reply Favorite

Date: January 25th, 2026 2:50 PM
Author: Chrome Whorehouse Private Investor

I use it to summarize depositions to clients then edit it, but you have to tell it to do a neutral analysis and point out strengths and weaknesses and it’ll be more balanced

(http://www.autoadmit.com/thread.php?thread_id=5826409&forum_id=2\u0026mark_id=5310744#49617676)



Reply Favorite

Date: January 25th, 2026 2:53 PM
Author: Racy Smoky Hospital

"In one of the convos, I framed the facts as the plaintiff. In the other, as the defendant. In both cases, ChatGPT confidently agreed that its side had the stronger legal position."

Did you ever consider framing the facts from both sides at the same time?

(http://www.autoadmit.com/thread.php?thread_id=5826409&forum_id=2\u0026mark_id=5310744#49617680)