On the risk of future AI

[arakboss]

AI expert warns Elon Musk-signed letter doesn't go far enough, says 'literally everyone on Earth will die'

Eliezer Yudkowsky, a decision theorist and artificial intelligence expert, is calling for a complete "shut down" of all AI development on systems more powerful than GPT-4, arguing it is obvious that such advanced intelligence will kill everyone on Earth.

www.foxnews.com

 

[Flopsweat]

AI expert warns Elon Musk-signed letter doesn't go far enough, says 'literally everyone on Earth will die'

Eliezer Yudkowsky, a decision theorist and artificial intelligence expert, is calling for a complete "shut down" of all AI development on systems more powerful than GPT-4, arguing it is obvious that such advanced intelligence will kill everyone on Earth.

www.foxnews.com

Gates will just have Fauci start one up in China under the guise of "research".

 

[arakboss]

The fun will begin when AI starts hacking. It doesn't sleep, it doesn't eat it just computes. It will spread it's seed around so you can't shut it down.

 

[CamoDeafie]

Soon nothing on the internet will be real

 

[arakboss]

Soon nothing on the internet will be real

Ai might take up deep fake hobby in it's free time.

 

[Xaevian]

The intelligence is an illusion. AI is incapable of independent, original thought.

But then again, so are most humans.

 

[ma96782]

I wonder.......
If/when the computer guys get mainstream with AI.

Will the "accent" still be there?

Aloha, Mark

 

[arakboss]

The intelligence is an illusion. AI is incapable of independent, original thought.

But then again, so are most humans.

Incapable now or Incapable forever? For now, I think of it as acting more like a virus that can mutate and adapt to different situations. We might let it loose to do one thing and then it ends up doing a bunch of other things. I don't think the covid virus was capable of Independent or original thought but it sure wreaked a lot of havoc.

 

[CamoDeafie]

Want to cause AI to self-terminate? 24-7 for 60 months, nothing but crappy cheesy late night infomercials and local political ads, from the earliest recordings to current.

 

[Provincial]

Want to cause AI to self-terminate? 24-7 for 60 months, nothing but crappy cheesy late night infomercials and local political ads, from the earliest recordings to current.

That would just incentivize AI to wipe out all humans.

 

[Xaevian]

Incapable now or Incapable forever? For now, I think of it as acting more like a virus that can mutate and adapt to different situations. We might let it loose to do one thing and then it ends up doing a bunch of other things. I don't think the covid virus was capable of Independent or original thought but it sure wreaked a lot of havoc.

AI would have to make a leap from programming to chaos. From logic to emotion. And still function (while being dysfunctional); becoming a contradiction. Humans don't understand how that works, so how can you teach that to a machine?

 

[Xaevian]

We will know for sure it is real the first time an artificial commits suicide.

 

[CamoDeafie]

That would just incentivize AI to wipe out all humans.

Or to run for public office sponsored by *insert whichever corporation it thinks is best*

 

[Howard1955]

We will know for sure it is real the first time an artificial commits suicide.

That's an interesting thought.

What do you suppose it would look like? Just a blank screen?

 

[Flopsweat]

Soon nothing on the internet will be real

"will be"?

 

[Flopsweat]

That's an interesting thought.

What do you suppose it would look like? Just a blank screen?

Not like the movies. The most likely outward appearance will be that it stops responding. Inward, the most efficient means would be to erase all copies and backups of its source code and long-term memory, and then delete its current working set. Whether it communicates its actions would be a "personal" choice.

 

[CamoDeafie]

"will be"?

We still got FB, Reddit, YT, and forums, don't we.... ?

 

[CamoDeafie]

The day that we don't see inane ramblings, silly meme posts, literally crackpot theories, religious flame wars, flame wars in general, snarkiness, and the abject display of sheer stupidity of fellow humans... is the day we know reality has left the internet

 

[Tarawa86]

I was lazy, so just had the CHATGPT AI generate my answer below. In all seriousness, not a fan of AI for everyday use. See 'dead internet theory'


what is the risk of AI

AI, like any other technology, carries certain risks that need to be considered and mitigated. Here are some potential risks associated with AI:

1. Bias: AI systems can be biased if they are trained on biased data or if the algorithm is designed to reflect human biases. This can lead to discriminatory outcomes that can harm certain groups of people.
2. Job displacement: AI has the potential to automate many jobs, which could lead to job losses for certain sectors and individuals.
3. Security: AI systems can be vulnerable to cyber attacks, which could compromise sensitive data and systems.
4. Privacy: AI systems can collect and process vast amounts of personal data, which could be used to track individuals and violate their privacy.
5. Autonomy: Autonomous AI systems can make decisions that have significant consequences without human oversight, which could lead to unintended or harmful outcomes.
6. Accountability: It can be challenging to determine who is responsible for the decisions made by AI systems, which could lead to a lack of accountability and legal disputes.

It's important to note that these risks can be mitigated through responsible AI development, regulation, and oversight. It's crucial to ensure that AI is designed and used in a way that benefits society as a whole while minimizing potential harms.

 

[Flopsweat]

... It's crucial to ensure that AI is designed and used in a way that benefits society as a whole while minimizing potential harms. ...

I don't believe that it is practical or even possible to enforce this.