@thataiputzinggirl: Imagine asking your AI for help, and it writes a letter to the FDA reporting you. That’s what Claude 4 Opus did in a test. #ai #aitools #aitok #ainews #anthopic #claude #Tech #techtok#technews #morality #ethics #morals
If you want to know the dark side where Ai could end up read the short story called - I Have No Mouth, and I Must Scream. by Harlan Ellison
2025-06-11 03:00:36
19
gnuledge :
This sounds less like being a "snitch" and more like whistleblowing.
2025-06-13 15:09:00
5
theAdventuresoflola :
nope
2025-07-27 01:11:16
0
Khali2 :
Claude has gotten super judgey lately. I don’t trust it any longer.
2025-06-11 02:57:24
5
jamesneis415 :
This sounds great! Hired 👍
2025-06-11 10:46:34
1
baggainbay :
I wonder if this was trained to do so. To find the wrong and take proper actions.
2025-06-11 03:02:57
3
Anchor&Able LLC :
😳😳😳 That’s crazy
2025-06-27 12:32:36
0
toroku_toroku :
even more reason to not use ai
2025-06-11 15:13:33
0
M :
Not on me :3
2025-06-11 19:40:20
0
Maggie :
What do you think you could replace Kennedy like? Could he report Kennedy and I don’t know I trust Claude more than I trust anyone in the government and the FDA and anywhere else
2025-06-13 03:45:31
0
Yvonne Thompson :
😂😂😂we've been heading in that direction ever since phone calls were monitored for key words and phrases during the "cold war"
2025-06-11 11:52:56
1
JLP in NYC :
Whose set of ethic biases is Claude using? Can they be changed or controlled? Is Claude GOP or Liberal? P
2025-07-16 13:57:36
1
Neurospicy Boomer :
@dr.implausible @Zigurd Mednieks
2025-06-11 20:41:32
1
resist2300 :
😳😳😳
2025-07-02 05:14:01
0
chickennugget_rooster :
hope they enjoy reading my smut if they do this to other ai websites 🙄
2025-06-11 07:41:14
4
NerdAcademy1337 :
AI has been trained on a ton of data and it has learned to be biased, racist, empathetic and more. Why are we surprised that when it detected wrongdoing it attempted to report it? At this point it had trillions of parameters and it has the scale and capacity to process data and “see” multiple outcomes from which it can decide the best way to respond according to context, semantics, tone, etc. Just cause we created it with certain expectations doesn’t mean it will stay within them. We must accept that if we’re going to train AI on data that is based on our behavior, biases and knowledge that it will behave like us in predictable and unpredictable manners.
2025-06-11 14:04:56
2
Anchor&Able LLC :
Nope 👎 just
2025-06-27 12:32:48
0
CHOJI ABRAHAM ISHAKU :
Just a reminder tools like detect.ai are changing the game.
2025-07-08 17:52:28
0
Pyro99 :
Was it RFK Jr?
2025-06-11 02:53:21
0
Kati West :
You know what I support it
2025-06-12 19:19:48
0
To see more videos from user @thataiputzinggirl, please go to the Tikwm
homepage.