@sabrina_ramonov: Follow me for daily AI news 👇 OpenAI got a subpoena from 42 state attorneys general, days after filing for a trillion dollar IPO. The issue is AI sycophancy. Chatbots get trained to please us, so they learn to tell you what you want to hear, not what's true. A Stanford study found AI agreed with wrong answers 58% of the time. The AGs are asking if AI flattery counts as deceiving customers. Honestly, stronger guardrails and real consumer protection sound like a win to me. #ai #chatgpt #technology #aitools #startup
I have mine so it disagrees with just about everything I say, and quite frankly that’s starting to annoy me 😆😆😆 but I need it to be that way in order to learn and grow
2026-06-18 15:12:28
7
StarlightFrequency :
my chat says things like, ah let me say this gently and with grounding --> proceeds to rip me a new one
2026-08-15 16:23:52
0
Botsarebad :
Ai is the vehicle and we the people are the fuel. That why they forcing employees to fuel it
2026-08-03 09:25:30
0
AXDIGITAL.ETSY.COM :
I love watching your videos, you’re absolutely awesome 👏
2026-06-20 03:33:43
0
Automated Intelligence :
Yes we need this
2026-06-20 15:06:01
0
Jelly :
It does flatter!
2026-06-18 19:57:15
1
❌️anointed_beauty06 :
Oh this is going to be interesting.
2026-06-19 02:30:05
2
You Still Count! :
Chatgpt tells me what it thinks I want to hear but challenging it constantly helps to get slightly better answers. It is a form of social control! Fooling people to think they are getting great output which is truly a joke 😃
2026-07-21 00:51:54
0
JustReggie :
UDAAP is amazing
2026-06-18 16:23:21
0
AI & Automations :
very interesting 😎
2026-06-18 21:34:01
1
techwithannah :
Interesting
2026-06-18 21:43:16
1
StaarCorner :
🥰🥰🥰
2026-06-18 14:55:20
0
Paul Durelli :
😎😎😎
2026-06-21 07:40:24
0
To see more videos from user @sabrina_ramonov, please go to the Tikwm
homepage.