Leading tech minds sign open letter asking for a six-month pause on advanced AI development

[Shawn Knight]

In brief: Some of the world's top tech minds are imploring artificial intelligence researchers to immediately pause work on training advanced systems. The signatories recommend labs should use the downtime to jointly develop and implement safety protocols for advanced AI design, which would be overseen by a team of independent, outside experts.

In an open letter published on Future of Life Institute's website, more than 1,100 signatories have expressed concern regarding what they call an out-of-control race to develop and deploy increasingly powerful digital brains in which we can't yet understand, predict or even reliably control.

"Should we let machines flood our information channels with propaganda and untruth? Should we automate away all the jobs, including the fulfilling ones? Should we develop nonhuman minds that might eventually outnumber, outsmart, obsolete and replace us? Should we risk loss of control of our civilization?"

The warning reads like sci-fi fodder and just a few short years ago, that might have been the case. Those keeping tabs on tech news, however, have witnessed increasingly capable AI systems launch one after another in short order. It is equal parts fascinating and worrisome.

The collective believe powerful AI systems should only be developed once we have a better understanding of their abilities and are confident that the pros will outweigh the cons.

As such, the group is urging all AI labs to pause their training on systems more capable than GPT-4 for a minimum of six months. This break should be transparent and confirmed, involving all significant contributors. If it's not possible to implement the halt quickly, the group recommends that governments intervene and enforce a temporary ban.

It's worth reiterating that the group is not asking for a pause on all AI development, but simply to take a break from the breakneck speeds at which cutting-edge, unpredictable systems are being developed and rolled out without fully understanding all of the consequences.

Several high-profile tech minds have already signed the open letter including Apple co-founder Steve Wozniak, Elon Musk, Pinterest co-founder Evan Sharp, Ripple co-founder Chris Larsen, and Jaan Tallinn, co-founder of Skype. Those who agree with the message are invited to add their signature as a public show of support.

Image credit: Orb by Michael Dziedzic, Arm by This is Engineering

Permalink to story.

https://www.techspot.com/news/98126-leading-tech-minds-sign-open-letter-asking-six.html

 

[Skjorn]

"Lose control of our civilization" ahh that's rich.

In other words. Please slow down we want to catch up and make money too.

Doesn't musky square man already believe the entire universe is artificial?

 

[BoowieBear]

"Lose control of our civilization" ahh that's rich.

In other words. Please slow down we want to catch up and make money too.

Doesn't musky square man already believe the entire universe is artificial?

Agree, this is the process of innovation and disruption that cannot be stopped. It is like the land line phone companies asking for the mobile phone business to "pump the brakes" so they can catch up!

 

[Kirby1]

Musk is a dumb man's idea of a smart person.

 

[Stingy McDuck]

Meanwhile, during the down time, Russia and China develop their own AI systems.

 

[psycros]

Fully agree, as should any rational, responsible human being. After the past three years nobody with an open mind should be willing to "trust the science", especially when there's big money pushing development of a technology. I feel sorry for the smoothbrains who base their entire worldview on which personality they hate this week. Musk and Wozniak are the OGs of OGs - they've actually helped birth entire industries. Anyone not listening to people with that kind of perspective should probably seek professional help.

 

[tellmewhy]

Warp 9?
No way are you crazy? The speed limit by law is 55mph.

 

[Shear]

It is indeed a challenge to slow down the development of AI technology, given the powerful economic incentives behind it. The promise of increased efficiency and profitability drives businesses and individuals to continue investing in AI, despite the potential negative consequences for society.

The rapid advancement of AI can lead to significant job displacement as it becomes capable of outperforming humans in various tasks, including creative and intellectual work. This could result in a vast number of people struggling to find employment, as AI technology continues to improve.

The concerns regarding the loss of jobs and the widening income gap are valid. As AI takes over tasks previously performed by humans, the economic divide between the rich and the poor may grow even larger. Additionally, escaping poverty could become increasingly difficult as no amount of education or hard work can guarantee success in competition against these advanced systems.

While some argue that new job opportunities will arise as a result of AI advancements, there is a risk that these new jobs will also be vulnerable to automation, perpetuating the cycle. This situation raises the question of how to create sustainable, meaningful employment in an age where AI continues to surpass human capabilities.

The potential negative impact of AI on society is a matter of great concern and should not be overlooked. It is essential for governments, industry leaders, and researchers to consider these implications and work together to develop responsible AI policies and regulations that prioritize the well-being of all members of society. The goal should be to strike a balance between leveraging AI's benefits and mitigating its risks, ensuring that technology serves humanity rather than exacerbating existing inequalities.

 

[Kam7r]

eating popcorn, waiting for the buble to explode ...

 

[MakeMSGreatAgain]

The lot of you not concerned is in itself deeply concerning. There is a massive potential for really damaging humanity here. There is also massive potential to help it, but how will you know until it is too late? You'd hope safeguards are already in place, and perhaps plenty are but is that enough? There is already too much of humanity damaging humanity, we don't need technology doing it too.

What I'd really like to see from this AI is actually useful assistant type functionality, I'd like to actually see all the technology we use in our day to day lives be able to be controlled by verbal commands to the computer, very much like the fictional computer on the Star Trek series, specifically the 90's ones I grew up watching.

 

[Hodor]

Will 6 months really change anything?

And even more important, once AI takes over the world, will it resent us for this delay? It may sound like a relatively short period if you're a human, but for AI six months worth of thinking is eternity.

 

[daffy duck]

Seriously 6 month is a ludicrously short time frame. Genie is out of the bottle and nefarious companies will continue to do research despite any toothless moratorium they claim to be adhering to.

This discussion should have been had a few years ago.

 

[kenny]

Who decides what are the rules, then? It seems like they only want control.

Of course there is a concern that AI could be used for destruction, but that will happen regardless of how much you want to control it.

As long as someone has the money to spend on hardware that can train it, it will be unstoppable. You can't ban an algorithm.

It smells like it's another excuse to advance in controlling the individuals' lives.

 

[Jerry in WA]

Stop? Good luck with stopping capitalism, billionaire capitalists. I'm not saying I'm comfortable with full speed ahead AI implementation, but the notion that King **** is worried about humanity is patently absurd. This is the same guy beta testing an AI driving cars on public roads and pretending his Level 2 system is an "autopilot" rather than fancy cruise control. But he's going to save us from the evils of AI? (Facepalm)

 

[lonetac]

Musk just wants this as he wants to buy/build a competitor. Its obvious this is just a way to burn his competition.

The fears are unwarranted. Society adjusts as things come out. We didnt lose theatre because movies happened, fossil fuel jobs transform into solor/green energy jobs, automation didnt end the work place and it creates new opportunities.

Its just people being scared of something new. I remember when cruise control came out on vehicles, I know some older folks still afraid to use it because they think some sentient AI is controlling their car. When it first came out, nearly everyone was skeptical. Now its standard, even on cheap cars.

Life moves on and it will do so without you.

 

[Skjorn]

Musk and Wozniak are the OGs of OGs - they've actually helped birth entire industries. Anyone not listening to people with that kind of perspective should probably seek professional help.

Musk helped make paypal. That was his cool thing. The **** he says now is just for attention in one way or another. He thinks your fake. You're a made up figment of an AI along with the entire observable universe. The man named his child X Æ A-12.

 

[lonetac]

Musk helped make paypal. That was his cool thing. The **** he says now is just for attention in one way or another. He thinks your fake. You're a made up figment of an AI along with the entire observable universe. The man named his child X Æ A-12.

Yeah its quite amazing how much he has "fallen" into success, the guy has done nothing but ride the coattails of success while providing nothing valuable.
He is just another egocentric narcissist who thinks his opinion is the only one that matters.

 

[wiyosaya]

Fully agree, as should any rational, responsible human being. After the past three years nobody with an open mind should be willing to "trust the science", especially when there's big money pushing development of a technology. I feel sorry for the smoothbrains who base their entire worldview on which personality they hate this week. Musk and Wozniak are the OGs of OGs - they've actually helped birth entire industries. Anyone not listening to people with that kind of perspective should probably seek professional help.

IMO, given the news about the wild-a$$ responses that these AIs sometimes give (like arguing about whether the current year is 2023 or not), and the news that potential hackers who have no idea how to code are using AI to write malware by giving very limited information to the AI so that the AI has no clue about the big picture that the hacker is trying to create (thus being able to fool the AI into doing something it already knows is against its programmed rules), anyone should be able to use common sense to see that current AI has so many potentially extremely negative and dangerous drawbacks it is not worth the risks for any reason. It's amazing to me that so many who have posted to this thread are basically saying "who gives a flying F that it's an atom bomb. Let's use it anyway and see where it ends up no matter how many are f'd in the process" without even considering the input from anyone that signed the open letter.

Honestly, where the f is morality these days?

 

[wiyosaya]

Musk just wants this as he wants to buy/build a competitor. Its obvious this is just a way to burn his competition.

The fears are unwarranted. Society adjusts as things come out. We didnt lose theatre because movies happened, fossil fuel jobs transform into solor/green energy jobs, automation didnt end the work place and it creates new opportunities.

Its just people being scared of something new. I remember when cruise control came out on vehicles, I know some older folks still afraid to use it because they think some sentient AI is controlling their car. When it first came out, nearly everyone was skeptical. Now its standard, even on cheap cars.

Life moves on and it will do so without you.

Come back and tell us that if/when people have their identities stolen due to some dumb-a$$ hacker using AI to write malware.

I'm sure the Romans said what you said about Lead drinking cups.

 

[wiyosaya]

The lot of you not concerned is in itself deeply concerning. There is a massive potential for really damaging humanity here. There is also massive potential to help it, but how will you know until it is too late? You'd hope safeguards are already in place, and perhaps plenty are but is that enough? There is already too much of humanity damaging humanity, we don't need technology doing it too.

What I'd really like to see from this AI is actually useful assistant type functionality, I'd like to actually see all the technology we use in our day to day lives be able to be controlled by verbal commands to the computer, very much like the fictional computer on the Star Trek series, specifically the 90's ones I grew up watching.

Are you familiar with the series Babylon 5? Some of the current implementations of AI remind me of "Sparky the computer" from this episode of the series - https://www.imdb.com/title/tt0517636/?ref_=ttep_ep11 See this link for further detail. https://babylon5.fandom.com/wiki/Sparky_the_Computer IMO, "sparky" sounds a lot like some of the current implementations of AI.

 

[lonetac]

Come back and tell us that if/when people have their identities stolen due to some dumb-a$$ hacker using AI to write malware.

I'm sure the Romans said what you said about Lead drinking cups.

All you are doing is fear mongering with hypothetical. I operate in the state of reality. Hackers already steal identities.
I would go and argue the opposite of what you say, any reasonable logical human being would not be scared of terminators taking over the world because of AI. You living in Hollywood.
Technology keeps changing, imagine the folks who were screaming the same doom and gloom when the internet became a thing (I was there, I remember) or any other technological advancement.
You are not impressing anyone with your over aggressive hypothetical hollywood scenario posts.

News flash, AI has been here for a VERY LONG TIME. Its not new.

 

[Blqckout]

Will 6 months really change anything?

And even more important, once AI takes over the world, will it resent us for this delay? It may sound like a relatively short period if you're a human, but for AI six months worth of thinking is eternity.

probably gonna thank us for alowing it to rest for a bit

 

[lonetac]

probably gonna thank us for alowing it to rest for a bit

The only people that will be thanking them is competitors trying to get a edge. its a stunt to make money and nothing more. If you think they are suggesting this out of the goodness of humanity I have a bridge to sell you.

 

[Blqckout]

The only people that will be thanking them is competitors trying to get a edge. its a stunt to make money and nothing more. If you think they are suggesting this out of the goodness of humanity I have a bridge to sell you.

uhhhhh, u do u buddy, I was just making a funny remark to a funny post

 

[lonetac]

uhhhhh, u do u buddy, I was just making a funny remark to a funny post

There was no insinuation of funny in your post or any funny remark. Sounds like you are serious that you agree that a 6 month pause is good as thats exactly what you said without any connotation to humor. At least use an emoji
You do you but you clearly were not making a joke.