[ad_1]
‘Near the singularity; unclear which side.’
This is the cryptic tweet sent by the architect of ChatGPT, Sam Altman, earlier this month that left the tech world pondering its meaning.
In his post on X, which was sent on January 4, it simply said: ‘I always wanted to write a six-word story. here it is.’
In a follow-up post moments later, he explained what the meaning of his short story is.
‘(It’s supposed to either be about 1. the simulation hypothesis or 2. the impossibility of knowing when the critical moment in the takeoff actually happens, but I like that it works in a lot of other ways too)’, he wrote.
For those of you who have been left even more puzzled, here it is in layman’s terms, at least as many understood it.
Altman, who is credited with helping expand AI, was warning that machines could soon surpass human intelligence, with far reaching and unknown consequences.
Around the same time, the OpenAI tech entrepreneur wrote on his personal blog that his company was ‘confident’ it now knew how to build artificial general intelligence (AGI), which is where a machine can do any intellectual task a human can do.
‘We are beginning to turn our aim beyond that, to superintelligence in the true sense of the word,’ he said.
This comes just ten years after one of the greatest minds of in a generation, Stephen Hawking, warned that forging ahead with this may not be a great idea.
The theoretical physicist, who died of Motor Neurone Disease in 2018, was an early adopter of AI. He used it to help him speak with a computerised voice.
But he had grave concerns about the future of the technology even before the buzz, and warned that humans could be engineering their own demise.
‘The development of full artificial intelligence could spell the end of the human race,’ he told the BBC.
‘It would take off on its own, and re-design itself at an ever increasing rate,’ he said.
‘Humans, who are limited by slow biological evolution, couldn’t compete, and would be superseded.’
Altman must have known that his comments would trigger furious speculation that we’ve already tipped into a future with super intelligent machines.
He later reined it in by saying: ‘Twitter hype is out of control again’ and ‘we are not gonna deploy AGI next month, nor have we built it’ (despite being the one to start the hype in the first place).
But with President Donald Trump recently announcing $500 billion of investment in AI infrastructure with The Stargate Project, a figure which makes the UK’s £14 billion bid to be a tech superpower look teeny, many think we really need to take Hawking’s warning seriously.
Geoffrey Hinton, known as the ‘Godfather of AI’ and a former Google executive who resigned his role to speak on the tech’s perceived dangers, said last month that the pace of AI was much faster than he had anticipated and required international treaties on the scale of those regulating nuclear weapons.
He told BBC Radio 4’s Today programme that he thought there was a ‘10% to 20%’ chance that AI would wipe out humans within the next three decades.
Dr Jonathan Aitken, who specialises in autonomous robotics at the University of Sheffield, told Metro the workings of powerful AIs are not easy to examine.
‘It’s difficult to know exactly the power of the model and the level of control that you give the model,’ he said.
‘As we develop more powerful AIs, it’s always difficult to know exactly what the realms are and how they will be used.’
But he said his main concern currently was not that AI itself is out of control, but that bad actors could corrupt and change the models.
To view this video please enable JavaScript, and consider upgrading to a web
browser that
supports HTML5
video
‘I’ve never gone in for the “AI will take over the world thing,” he said. ‘But we’re giving a lot of power to these models to make decisions.
‘If you can poison and change the model, you can change what people do.’
Generative AI, which can create convincing false videos, audio and social media posts (and has already flooded sites like Facebook and X with ‘slop’), could prove dangerous, he said. ‘If you put enough information in front of someone, you can change someone’s opinion.
‘You can change behaviour in the real world, and that’s the real danger of this, rather than a piece of AI with its tentacles everywhere.
‘My concern isn’t that AI will end the world, but the actions of an AI system will cause people to do something stupid.’
Artist and musician Laurie Anderson, who uses AI in her work, also raised this as a concern recently, when she told Desert Island discs that she nevertheless believes it is an ‘end of the world kind of thing’ due to how easy it now is to impersonate anyone with fake videos and audio.
The widow of Lou Reid said: ‘You could win an election with something like this. You could start a war with something like this.
‘We depend on a certain amount of authenticity that gets stamped, but we don’t know who’s saying what anymore at all.’
So is the toothpaste out of the tube? Will our own microwaves grow legs and fry us?
That’s probably not how the robot revolution would unfold… although who really knows?
Hinton said the problem would be that our capabilities would just be vastly surpassed by AI, like a toddler compared to its parents.
While a parent is generally happy to act in service to their child, there are few other examples of a more intelligent being allowing itself to be controlled.
Such a milestone, when machines no longer need human input to keep improving, has generally been thought to be decades away, if possible at all.
A comparison often made is to the event horizon of a black hole, which offers no way of return once crossed and will lead to unpredictable changes where rules no longer apply.
You wouldn’t see anything change immediately –unless you tried to get back out.
With money pouring into AI from governments around the world, it’s unlikely we’ll be reversing soon in any case.
Get in touch with our news team by emailing us at [email protected].
For more stories like this, check our news page.
[ad_2]
Source link
