I note that the big owners of AI programs, such as Meta, say the fears are overblown. But they would, wouldn’t they!
In order to truly create public benefit, we need mechanisms of accountability. The world needs a generative AI global governance body to solve these social, economic, and political disruptions beyond what any individual government is capable of,
[This article you may need the paywall ladder to jump over.]
I firmly do believe an independent AI global governing body must be formed ASAP and a pause in AI development be enforced until the governing body is up and running.
Regarding the tech industry being left to self censor,
When an industry experiences heightened scrutiny, barring off prohibitive regulation often means taking control of a narrative – ie calling for regulation, while simultaneously spending millions in lobbying to prevent the passing of regulatory laws.
Does not say much when the BBC article quote’s UK PM Sunak
"Now that’s why I met last week with CEOs of major AI companies to discuss what are the guardrails that we need to put in place, what’s the type of regulation that should be put in place to keep us safe.”
When I was younger I read a SF book describing man’s attempts to create a computer so powerful that it could be asked the ultimate question. “Is there a god?”. The answer çame back- “There is now.” Are we approaching that position?
I expect that was ‘Answer’ by Frederic Brown, an early sci-fi writer. A bit chilling.
I doubt we will ever reach the point where a machine will be that powerful or omnipotent over mankind. The greater danger is of a human using the power of machines to lord it over all other humans. I’m hoping that there will be fail safes built in to prevent a single source control.
But Google Earth requires cars with cameras driving all over the world. In just a few seconds, this tool has done a surprisingly good job of creating what the houses where I live look like… I mean, this isn’t where I live, but it’s got the style of buildings in my commune pretty well.
Plus, whereas Google can only update Google Earth photos once every x number of years, I could presumably tell this tool to recreate the same image but on a snowy winter’s day or in the early hours of a day, etc… It’s not going to set the world on fire but it’s been a fun distraction for a few minutes.
I would in all honesty not be surprised that military use of AI disables or circumvents the primary, yet fictional, Asimov ‘rules’.
The problem lies in how to direct ‘protect’ all humanity, in preference to one person It would not be unusual within the military to accept that a single person, say an operator, would not be indispensable.
What went awry in the drone simulation is not that the AI program ‘decided’ autonomously to kill the operator but that there were no controls on whom it could and who not kill.
The greater risk to humanity now is an old-fashioned one: Humans creates an AI with a goal of acquiring superior power, and it succeeds.
As I think he’s retired and was an invited speaker, ''mis-spoke might mean he just dramatically embellished his presentation , or in old-fashioned terms, he simply ‘lied’. Remember the days when lies were reprehensible? Or have they always been present, but today it no longer matters?
What may be pertinent is his position as Director of the Department of the Air Force and MIT AI Accelerator, Cambridge, MA.
The anecdote in his speech may have been to spice up events but I suspect the simulation did occur in the way he reported, bein a man in the inside he would know, but the Department doesn’t like the public being made aware of such details. “Misspoke” is a way of saying his report of a classified event was not permitted, not that he lied.
(You can see I’m concerned enough to fact check the progress of AI on a daily basis! )
I remember reading something written by Herodotus from his travels around the then known world, describing the things he’d seen in various places like Egypt, then moving seamlessly to describe fantastic animals from another part of the world that could not possibly have existed with no obvious change in writing style (possibly due to the translation). My conclusion was that Facebook etc had only revealed what was already in the human heart, of a desire to make things up regardless of reality.
Aaahhh, yes! Bless. Having lived many years and travels through Asia, I learnt there are cultural reasons for giving information, or even answering direct questions, to someone in the way they are thought to want.
Thus, “Do you serve breakfast?” could result in your thinking breakfast might appear. We learnt to never ask ‘closed questions’ (where the answer could be yes or no) because in an effort to please us the answer would invariably be “Yes!” although the reality was in fact no. So we would instead ask “What time do you serve breakfast?” The ensuing look of puzzlement would provides an answer. No lies.
Travellers’ tales endured and indeed probably grew in subsequent eras.
The oldest artwork in our inconveniently vast collection is a lovely woodcut illustration in a page from Sebastian Münster’s1544 Cosmographia, the first European encyclopaedia. The book was presumably vandalised at some point in the past, which was probably a bad idea economically as well as culturally. Anyhow, our print depicts an Indian elephant and on its back is a small(ish) but implausible crenelated fort containing a dozen or more fierce turbaned and moustachioed Indian warriors.
One of the most interesting of these travellers’ tales is that of the S American sloth. Spanish artists accompanied early colonial voyages and accurately recorded the sloth hanging upside down from branches. However when the woodcuts were printed in early encyclopaedias the sloths were reversed as the printers assumed they correcting some error and so the animals were depicted walking on top of the branches. For some unrelated and unknown reason, many also were given humanoid faces…