General
Share and discuss tech, products, business, startups, or product recommendations
Start new thread
Maria Noemi Hernandez

Can you recognize an AI-generated response? How do you feel about them?

by
Collato
•
Without knowing it, I became a detective at spotting them on discussions, product launches, Twitter, Linkedin, and different platforms.
33 views

Add a comment

Replies

Best
Ansar Uddin Khan
As far as I know , nowadays there are certain algorithms that are giving the human touch to the ai itself. So recognition of AI responses is quite difficult.
Maria Noemi Hernandez
@ansar_uddin_khan yes, as with everything in life... it depends. It depends if they're editing the answer after, how good the prompt is, how good the tool is, if they are posting what they get, etc
Catherine Norris
I definitely think I can recognize them now...but it's getting trickier and trickier!
Maria Noemi Hernandez
@catherine_norris Every day more and more AI tools are coming out that make the answers better and better. For me, it is a personal challenge now😂
Ra De
Rade Wade
Irina Haupt
Sometimes yes, sometimes no. Long, elaborated, and somewhat too precise responses are easy to spot, but then there are some mechanisms where you can adjust the tone (sarcasm, emotional, etc.) and this makes answers sound more human and so much more difficult to spot. 🧟 but I love how PH and other tools have already implemented the Report as spam/AI-generated response :) :)
Maria Noemi Hernandez
@irina_haupt I like the personal messages, and they are easy to spot. You're right with so many new options to personalize the response, it is becoming more and more difficult.
Elly
It depends on how the AI was trained, but in many cases, I think I can still spot it. 70% chance? haha 🤓
Maria Noemi Hernandez
@elly_cho That's still a lot!
Elly
@mimih_arteaga haha 😁
Ra De
@elly_cho 😁😁😁😁
Lisa Kim
I honestly don't know. These days, I assume that a perfectly crafted paragraph for a small social media post is ChatGPT
Hello_Joy
@mycelebs_elle I'm not sure either :0
Maria Noemi Hernandez
@mycelebs_elle Same here! And I think I also recognize some emojis👀
David Cagigas
If it has been worked on after its generated then it's hard to spot. But yeah otherwise you can spot it! I think it's acceptable where the tasks are repetitive or of not great importance. But otherwise it can be a bit tricky.
Maria Noemi Hernandez
@edworking I think it's nice to keep it fresh and personal for discussions, forums, or conversations on Social Media. I like to read what the other person is thinking and not AI. But I agree that there are some tasks that AI is making so much easier!
Charlie Kor
If someone uses AI to proofread and rephrase a comment, I may not be able to identify it.
Maria Noemi Hernandez
@charlie_lee1 same here! If it is edited by a human and brings a more personal touch, then I think it's ok. Ultimately, we all need ways to improve and speed up our work.
Anna
Sure, there's a certain structure that the response has. Plus there are lots of generalizations
Maria Noemi Hernandez
@anya1 I know, you can see the exact keywords. A trick I recently learned is to improve the prompts; that way, you will get a more natural and less AI answer.
May
@anya1 @mimih_arteaga that’s curious! Could you share an example of how a prompt can be improved to get more natural answers?
Saeed Samir
I think I can recognize most of them
Maria Noemi Hernandez
@saeed_samir welcome to the team
Grace Hur
Not so much - I'm quite a gullible person 🙈.
Maria Noemi Hernandez
@gracehur hahahah that's ok too!
12
Next
Last