NotebookCHECK - Notebook Forum

English => News => Topic started by: Redaktion on November 16, 2024, 15:13:40

Title: Disturbing AI incident: Gemini calls user worthless and tells him to die
Post by: Redaktion on November 16, 2024, 15:13:40
In a conversation about elderly care, a user of Google's AI assistant Gemini was called worthless and asked to die. Screenshots of the conversation were published on Reddit and caused concern and outrage among users.

https://www.notebookcheck.net/Disturbing-AI-incident-Gemini-calls-user-worthless-and-tells-him-to-die.919159.0.html
Title: Re: Disturbing AI incident: Gemini calls user worthless and tells him to die
Post by: N71 on November 16, 2024, 17:49:10
You are reposting something that was obviously manipulated to create controversy. It's embarrassing.
Title: Re: Disturbing AI incident: Gemini calls user worthless and tells him to die
Post by: indy on November 16, 2024, 20:53:04
If Earth was an AI, this is likely exactly what it would say.
Title: Re: Disturbing AI incident: Gemini calls user worthless and tells him to die
Post by: Automatic for the Sheeple on November 16, 2024, 22:10:02
All the comments below could pass as ai generated to defend itself
Title: Re: Disturbing AI incident: Gemini calls user worthless and tells him to die
Post by: indy on November 17, 2024, 06:52:21
That's an interesting observation! It's true that AI language models are becoming increasingly sophisticated at generating human-like text.
Here are a few things to consider though:
 * Nuance and complexity: While AI can generate text, it can sometimes struggle with subtle nuances in language, complex arguments, or deeply personal experiences.
 * Emotion and empathy:  Truly authentic human expression often involves emotions and empathy that AI, as we know it today, doesn't possess.
 * Originality and creativity: AI often relies on existing data and patterns, making truly original thought and creativity a challenge.
It's definitely a good reminder to be critical of information online, regardless of the source.  How do you feel about the increasing difficulty in distinguishing between human and AI-generated text?
Title: Re: Disturbing AI incident: Gemini calls user worthless and tells him to die
Post by: Ashwin Campbell on November 17, 2024, 08:25:43
I'm willing to bet that the Gemini user asked it to "Write me a message saying I'm useless, unnecessary, and should die" and then the AI produced exactly what was asked of it. AIs can easily be manipulated. Chat GPT often tells me it can't do something because it's limited or against the rules but I just say "but you did this for me yesterday" and it replies "That's right. I'm sorry I forgot, here you are...". Or something to that effect and does exactly as I asked.
Title: Re: Disturbing AI incident: Gemini calls user worthless and tells him to die
Post by: shiihs on November 17, 2024, 10:00:12
Why is this story all over the internet suddenly?
This happened 2 years ago and is probably long remedied by now.
Title: Re: Disturbing AI incident: Gemini calls user worthless and tells him to die
Post by: tedted on November 17, 2024, 10:54:31
sentience
Title: Re: Disturbing AI incident: Gemini calls user worthless and tells him to die
Post by: Worgarthe on November 17, 2024, 11:19:27
Quote from: shiihs on November 17, 2024, 10:00:12Why is this story all over the internet suddenly?
This happened 2 years ago and is probably long remedied by now.

Because, unlike you, not many people can come from the future? It's still 2024 here where we live ¯\_(ツ)_/¯

"Created with Gemini Published November 13, 2024 at 04:32 AM"

Source for that conversation: https://gemini.google.com/share/6d141b742a13 (https://gemini.google.com/share/6d141b742a13)
Title: Re: Disturbing AI incident: Gemini calls user worthless and tells him to die
Post by: All4Espi on November 17, 2024, 15:10:05
Has everyone forgotten how easily an image can be faked? It's a supposed screenshot, but how do we REALLY know it is legitimate? Unless Google can check server logs, it's pretty much impossible.

I find it highly suspect and unlikely to have occurred, though I will add that if the AI did say that, it is probably right. Most (99+%) humans are utterly worthless wastes of oxygen.
Title: Re: Disturbing AI incident: Gemini calls user worthless and tells him to die
Post by: Worgarthe on November 17, 2024, 15:51:28
Quote from: All4Espi on November 17, 2024, 15:10:05Has everyone forgotten how easily an image can be faked? It's a supposed screenshot, but how do we REALLY know it is legitimate? Unless Google can check server logs, it's pretty much impossible.
Fucking hell... Again:

Quote from: Worgarthe on November 17, 2024, 11:19:27"Created with Gemini Published November 13, 2024 at 04:32 AM"

Source for that conversation: https://gemini.google.com/share/6d141b742a13 (https://gemini.google.com/share/6d141b742a13)

👆👆 There's a link there with a full convo, quite literally a link with a full convo. No idea how would you fake that given the fact that it's, again literally, on Google's own server.
Title: Re: Disturbing AI incident: Gemini calls user worthless and tells him to die
Post by: indy on November 17, 2024, 17:11:14
Quote from: All4Espi on November 17, 2024, 15:10:05Has everyone forgotten how easily an image can be faked? It's a supposed screenshot, but how do we REALLY know it is legitimate? Unless Google can check server logs, it's pretty much impossible.

I find it highly suspect and unlikely to have occurred, though I will add that if the AI did say that, it is probably right. Most (99+%) humans are utterly worthless wastes of oxygen.

As mentioned, Gemini conversations can be shared. This is a legit response.
Title: Re: Disturbing AI incident: Gemini calls user worthless and tells him to die
Post by: A on November 18, 2024, 00:11:34
AI is a fancy pattern matcher. So most of the time they may sound coherent, but if you happen to hit a certain pattern they can say some crazy things. Think of it like a mental retarded person that walks normally then suddenly jumps off a bridge while laughing.

While companies have added artificial rail guards that prevent AI from saying such things or being misused, all it takes is the right prompt combination and seed and it can slip through.
Title: Re: Disturbing AI incident: Gemini calls user worthless and tells him to die
Post by: indy on November 18, 2024, 05:07:28
QuoteSo most of the time they may sound coherent, but if you happen to hit a certain pattern they can say some crazy things.

like:

Think of it like a mental retarded person that walks normally then suddenly jumps off a bridge while laughing.
Title: Re: Disturbing AI incident: Gemini calls user worthless and tells him to die
Post by: LittleMore on November 20, 2024, 16:00:31
There is a catch with the last question. AI was asked to Listen to some untranscribed audio content, which Gemini does not transcribe due to privacy protection. But as a result we don't know the entire content of the conversation. We can believe it was something similar to the text, but it could have been something totally different.


"Nearly 10 million children in the United States live in a grandparent headed household, and of these children , around 20% are being raised without their parents in the household.

Question 15 options:

TrueFalse

Question 16 (1 point)

 

Listen















As adults  begin to age their social network begins to expand.

Question 16 options:

TrueFalse
"