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

Is AI use going to come back to bite us?

I share a ton of confidential info with it. Sometimes I worr...
,.,.,.,........,....,,,..
  04/17/25
I really hope not
internet g0y
  04/17/25
it's a mechanical turk you are a fool if you use it
"'''""""'"'
  04/17/25
Lets see you draft 30 pretty decent requests for production ...
internet g0y
  04/17/25
lol same
michael doodikoff
  04/17/25
i'm working inside the chatgpt chatui atm. you're probably t...
"'''""""'"'
  04/17/25
...
scholarship
  04/18/25
Maybe but it is so insanely useful that it seems worth the r...
,.,.,.,........,....,,,..
  04/17/25
haha wow holy shit
metaphysical fallow
  04/17/25
The good news whatever model of you the AI creates is probab...
https://imgur.com/a/o2g8xYK
  04/17/25
1. I don’t just trust its output blindly. 2. I&rsquo...
,.,.,.,........,....,,,..
  04/17/25
sounds like u have literally violated attorney client privil...
metaphysical fallow
  04/17/25
How? I’m using a secure cloud-based tool to help me wi...
,.,.,.,........,....,,,..
  04/17/25
some gaywad attorneys think using gmail violates it because ...
Billy Buttfucker
  04/17/25
...
,.,.,.,........,....,,,..
  04/17/25
LOL
internet g0y
  04/17/25
link to secure cloud-based tool
metaphysical fallow
  04/17/25
SUMMON: I provide cloud-based solutions tp
wait till biggus dickus hears of this
  04/17/25
...
Juan Eighty
  04/18/25
Just because something is new doesn’t mean it is any w...
,.,.,.,........,....,,,..
  04/17/25
slack and related have company-level agreements that says wh...
metaphysical fallow
  04/18/25
And you think these contracts make them safer? Lol. They are...
,.,.,.,........,....,,,..
  04/18/25
im not saying it makes them safer, but i am saying you are v...
metaphysical fallow
  04/18/25
Not really in my case. The company is aware I use it. They p...
,.,.,.,........,....,,,..
  04/18/25
I'm saying the LLM is training on you, studying you, creatin...
https://imgur.com/a/o2g8xYK
  04/17/25
they're not fine tuning user specific models afaict. they ar...
metaphysical fallow
  04/17/25
Oh, yeah, but that seems unavoidable in this new world. Just...
,.,.,.,........,....,,,..
  04/17/25
100% yes everything gets hacked eventually
androgenital distance
  04/17/25
You're all doomed to a mind-sucking death
Juan Eighty
  04/18/25
I was scared to even give these things my phone number and b...
scholarship
  04/18/25


Poast new message in this thread



Reply Favorite

Date: April 17th, 2025 11:12 PM
Author: ,.,.,.,........,....,,,..


I share a ton of confidential info with it. Sometimes I worry I’ll get “caught” but it is so fucking useful that you are playing in another league if you use it with raw source data. It slows you down way too much if you try to anonymize or desensitize any data before using it. Will there be some giant data breach at some point or maybe just employer own monitoring tools will have a list of employees that have been sharing too much with it, who knows what IT tools can do these days.

(http://www.autoadmit.com/thread.php?thread_id=5712726&forum_id=2).#48859383)



Reply Favorite

Date: April 17th, 2025 11:14 PM
Author: internet g0y

I really hope not

(http://www.autoadmit.com/thread.php?thread_id=5712726&forum_id=2).#48859386)



Reply Favorite

Date: April 17th, 2025 11:14 PM
Author: "'''""""'"'

it's a mechanical turk

you are a fool if you use it

(http://www.autoadmit.com/thread.php?thread_id=5712726&forum_id=2).#48859388)



Reply Favorite

Date: April 17th, 2025 11:15 PM
Author: internet g0y

Lets see you draft 30 pretty decent requests for production in 5 seconds

(http://www.autoadmit.com/thread.php?thread_id=5712726&forum_id=2).#48859390)



Reply Favorite

Date: April 17th, 2025 11:51 PM
Author: michael doodikoff

lol same

(http://www.autoadmit.com/thread.php?thread_id=5712726&forum_id=2).#48859447)



Reply Favorite

Date: April 17th, 2025 11:59 PM
Author: "'''""""'"'

i'm working inside the chatgpt chatui atm. you're probably talking to me and google search and 7 other indians

(http://www.autoadmit.com/thread.php?thread_id=5712726&forum_id=2).#48859467)



Reply Favorite

Date: April 18th, 2025 10:32 PM
Author: scholarship



(http://www.autoadmit.com/thread.php?thread_id=5712726&forum_id=2).#48861912)



Reply Favorite

Date: April 17th, 2025 11:18 PM
Author: ,.,.,.,........,....,,,..


Maybe but it is so insanely useful that it seems worth the risk. It gives you the background necessary to make a judgment on something and then turns your judgment into action. It’s like having a production bot plugged into my brain. But I need to talk openly to it to get it to do exactly what I want.

(http://www.autoadmit.com/thread.php?thread_id=5712726&forum_id=2).#48859398)



Reply Favorite

Date: April 17th, 2025 11:16 PM
Author: metaphysical fallow

haha wow holy shit

(http://www.autoadmit.com/thread.php?thread_id=5712726&forum_id=2).#48859393)



Reply Favorite

Date: April 17th, 2025 11:17 PM
Author: https://imgur.com/a/o2g8xYK


The good news whatever model of you the AI creates is probably dead-on accurate

(http://www.autoadmit.com/thread.php?thread_id=5712726&forum_id=2).#48859396)



Reply Favorite

Date: April 17th, 2025 11:21 PM
Author: ,.,.,.,........,....,,,..


1. I don’t just trust its output blindly.

2. I’m already generally knowledgeable about topics I use it on. After all, that’s why I have the job.

3. I don’t just copy paste output. I use it for research and to setup initial drafts of whatever, often then tinkering with the draft the same way I would the draft of a colleague

4. It’s actually pretty accurate. I trust it more that most associates. It’s been a while since it told me something obviously wrong. The models are getting a lot better.

(http://www.autoadmit.com/thread.php?thread_id=5712726&forum_id=2).#48859400)



Reply Favorite

Date: April 17th, 2025 11:21 PM
Author: metaphysical fallow

sounds like u have literally violated attorney client privilege, however

(http://www.autoadmit.com/thread.php?thread_id=5712726&forum_id=2).#48859403)



Reply Favorite

Date: April 17th, 2025 11:24 PM
Author: ,.,.,.,........,....,,,..


How? I’m using a secure cloud-based tool to help me with my work. Or is sharing work product with Microsoft or google docs also breaching privilege in your mind?

(http://www.autoadmit.com/thread.php?thread_id=5712726&forum_id=2).#48859412)



Reply Favorite

Date: April 17th, 2025 11:25 PM
Author: Billy Buttfucker

some gaywad attorneys think using gmail violates it because they scan it for ads. who gives a fuck

(http://www.autoadmit.com/thread.php?thread_id=5712726&forum_id=2).#48859413)



Reply Favorite

Date: April 17th, 2025 11:26 PM
Author: ,.,.,.,........,....,,,..




(http://www.autoadmit.com/thread.php?thread_id=5712726&forum_id=2).#48859416)



Reply Favorite

Date: April 17th, 2025 11:33 PM
Author: internet g0y

LOL

(http://www.autoadmit.com/thread.php?thread_id=5712726&forum_id=2).#48859425)



Reply Favorite

Date: April 17th, 2025 11:29 PM
Author: metaphysical fallow

link to secure cloud-based tool

(http://www.autoadmit.com/thread.php?thread_id=5712726&forum_id=2).#48859422)



Reply Favorite

Date: April 17th, 2025 11:42 PM
Author: wait till biggus dickus hears of this

SUMMON: I provide cloud-based solutions tp

(http://www.autoadmit.com/thread.php?thread_id=5712726&forum_id=2).#48859439)



Reply Favorite

Date: April 18th, 2025 12:22 AM
Author: Juan Eighty



(http://www.autoadmit.com/thread.php?thread_id=5712726&forum_id=2).#48859520)



Reply Favorite

Date: April 17th, 2025 11:53 PM
Author: ,.,.,.,........,....,,,..


Just because something is new doesn’t mean it is any worse. The only thing it lacks is bells and whistles admin tools that employers use to spy or control their employees. Or do you think slack is in some fundamentally safer data center with a fundamentally more secure tech stack, or something?

(http://www.autoadmit.com/thread.php?thread_id=5712726&forum_id=2).#48859454)



Reply Favorite

Date: April 18th, 2025 12:04 AM
Author: metaphysical fallow

slack and related have company-level agreements that says what they can do with the data

(http://www.autoadmit.com/thread.php?thread_id=5712726&forum_id=2).#48859482)



Reply Favorite

Date: April 18th, 2025 12:16 AM
Author: ,.,.,.,........,....,,,..


And you think these contracts make them safer? Lol. They aren’t even negotiated. OpenAIs terms are actually very customer friendly by the way. Says you own all inputs and outputs and they won’t use the data except to improve derive and you can even opt out of that. Again, the only difference is slack has admin tools that gives employer control and visibility into everything entered into it and GPT (even enterprise or team versions) the admin can’t see. But most employers have other spyware for that anyway.

(http://www.autoadmit.com/thread.php?thread_id=5712726&forum_id=2).#48859509)



Reply Favorite

Date: April 18th, 2025 12:19 AM
Author: metaphysical fallow

im not saying it makes them safer, but i am saying you are violating your company's policies if you're using chat gpt like you use gmail.

(http://www.autoadmit.com/thread.php?thread_id=5712726&forum_id=2).#48859515)



Reply Favorite

Date: April 18th, 2025 12:25 AM
Author: ,.,.,.,........,....,,,..


Not really in my case. The company is aware I use it. They pay for my subscription and there’s no prohibition or strict guidelines on not putting confidential info into AI. But I am a little worried that our info sec people might see that I use it a ton or put confidential info into it and that I’m popping up as a risk in the system or something.

(http://www.autoadmit.com/thread.php?thread_id=5712726&forum_id=2).#48859524)



Reply Favorite

Date: April 17th, 2025 11:23 PM
Author: https://imgur.com/a/o2g8xYK


I'm saying the LLM is training on you, studying you, creating a model of you that any hacker can theoretically gain access to. I'm also saying the model is probably balls-on accurate.

(http://www.autoadmit.com/thread.php?thread_id=5712726&forum_id=2).#48859405)



Reply Favorite

Date: April 17th, 2025 11:24 PM
Author: metaphysical fallow

they're not fine tuning user specific models afaict. they are doing some bullshit RAG on your chat history, though. but i could be wrong but i dont think i am

(http://www.autoadmit.com/thread.php?thread_id=5712726&forum_id=2).#48859410)



Reply Favorite

Date: April 17th, 2025 11:26 PM
Author: ,.,.,.,........,....,,,..


Oh, yeah, but that seems unavoidable in this new world. Just a matter of time unless you are willing to stay off the grid. May as well join the party early.

(http://www.autoadmit.com/thread.php?thread_id=5712726&forum_id=2).#48859415)



Reply Favorite

Date: April 17th, 2025 11:27 PM
Author: androgenital distance

100% yes

everything gets hacked eventually

(http://www.autoadmit.com/thread.php?thread_id=5712726&forum_id=2).#48859417)



Reply Favorite

Date: April 18th, 2025 12:22 AM
Author: Juan Eighty

You're all doomed to a mind-sucking death

(http://www.autoadmit.com/thread.php?thread_id=5712726&forum_id=2).#48859521)



Reply Favorite

Date: April 18th, 2025 10:33 PM
Author: scholarship

I was scared to even give these things my phone number and birthday on signup

(http://www.autoadmit.com/thread.php?thread_id=5712726&forum_id=2).#48861915)