Slashdot Mirror
What Does Artificial Intelligence Actually Mean? (qz.com)
An anonymous reader writes: A new bill (pdf) drafted by senator Maria Cantwell asks the Department of Commerce to establish a committee on artificial intelligence to advise the federal government on how AI should be implemented and regulated. Passing of the bill would trigger a process in which the secretary of commerce would be required to release guidelines for legislation of AI within a year and a half. As with any legislation, the proposed bill defines key terms. In this, we have a look at how the federal government might one day classify artificial intelligence. Here are the five definitions given:
A) Any artificial systems that perform tasks under varying and unpredictable circumstances, without significant human oversight, or that can learn from their experience and improve their performance. Such systems may be developed in computer software, physical hardware, or other contexts not yet contemplated. They may solve tasks requiring human-like perception, cognition, planning, learning, communication, or physical action. In general, the more human-like the system within the context of its tasks, the more it can be said to use artificial intelligence.
B) Systems that think like humans, such as cognitive architectures and neural networks.
C) Systems that act like humans, such as systems that can pass the Turing test or other comparable test via natural language processing, knowledge representation, automated reasoning, and learning.
D) A set of techniques, including machine learning, that seek to approximate some cognitive task.
E) Systems that act rationally, such as intelligent software agents and embodied robots that achieve goals via perception, planning, reasoning, learning, communicating, decision-making, and acting.
A) Any artificial systems that perform tasks under varying and unpredictable circumstances, without significant human oversight, or that can learn from their experience and improve their performance. Such systems may be developed in computer software, physical hardware, or other contexts not yet contemplated. They may solve tasks requiring human-like perception, cognition, planning, learning, communication, or physical action. In general, the more human-like the system within the context of its tasks, the more it can be said to use artificial intelligence.
B) Systems that think like humans, such as cognitive architectures and neural networks.
C) Systems that act like humans, such as systems that can pass the Turing test or other comparable test via natural language processing, knowledge representation, automated reasoning, and learning.
D) A set of techniques, including machine learning, that seek to approximate some cognitive task.
E) Systems that act rationally, such as intelligent software agents and embodied robots that achieve goals via perception, planning, reasoning, learning, communicating, decision-making, and acting.
Those look like legally workable definitions, (though I imagine I'd ultimately be proven wrong by billions of dollars' worth of tedious court cases).
The criterion for artificial intelligence used here doesn't really differentiate much from general computer use.
Not unlike the intent of the three laws of robotics, it would be wise beyond the normal abilities of those in government to get a sensible definition and implementation of rules in place before corporations have a billion dollar interest in the outcome.
Happiness in intelligent people is the rarest thing I know.
Ernest Hemingway
A big computer for other projects that can do advanced parlor tricks when needed for visiting dignitaries.
All the perception, planning, reasoning, learning is done by humans who need to get more funding.
AI is social engineering.
Domestic spying is now "Benign Information Gathering"
These are way too broad for a workable regulation. Anything that has a neural network could be regulated and that is just too rudimentary a technology to be usefully regulated. Also, it's probable that even things like adaptive filters could be regulated under such a definition. If the bill regulates the areas in which the applications are used - e.g. driving vehicles, surveillance, etc. where the federal government already has an interest, well MAYBE that's OK, but this seems like an easy overreach nonetheless.
"Thou shalt not make a machine in the likeness of a human mind."
To begin with, referring to "human intelligence" is pointless as we do not agree on what this is. Including "rational thinking" as part of the definition won't help either since the process of asserting rationality is non trivial. "To think as humans" and say that all artificial neural networks does this is to insult neurologists. It might work to say that neural networks are loosely inspired by how we think human brains work.
However, I do like if the definition include a metric for how the system can adapt and learn. Just don't confuse the two. Neither adaptation nor learning says anything if the system is about to converge to a stated policy. Sometimes you don't want adaptation but forceful behavior. Sometimes you realize your policy is crap but you don't want the system to cheat anyway. So are these attributes even useful to consider when you want to regulate AI (or human intelligence for that matter).
Why would the regulation of AI be any different from the regulation of human intelligence? Let the AI prove itself, then hold it or its "parents" accountable.
Back in high school in my computer programming class, we were taught arrays, to do this we made the game of memory where we had 16 cards with 8 matching values, which were randomized.
Then we were to pick 2 cards if we got a match we would had got a point. Then the computer picked two cards.
Normally most of the students just had the computer pick randomly. I felt ambitious as programming was my thing that made me the Alpha geek back then. So I made it keep track of the cards when it found them and learned from its mistakes, thus being a difficult game to play.
This isn't AI, but it seems to fit definition A. as would most video games of any challenge. Also most business intelligence apps that find patterns would classify.
If something is so important that you feel the need to post it on the internet... It probably isn't that important.
It's a series of tubes OK, that's all you need to know.
we should just ask Alexa.
I can't, I'll have to get my kids or my wife to ask Alexa. Alexa still doesn't recognise a British accent (at least not in the US). If I ask Alexa what AI is, I'll probably get a pizza delivered.
"That's the way to do it" - Punch
So let's discuss the why before we just start regulating stuff that 99.999% of the time will not need any regulation for any public safety, or even ethical purpose.
What is the purpose of regulating computer software? AI in most cases these days means computer software that has been trained with examples to process a data set rather than programmed to process one. It is just more efficient than figuring it out and programming an algorithm directly for more variable input. And once the training is over and optimized the algorithm is usually frozen so that it can be applied in a tested and predictable way. So AI is rarely about algorithms that are trained during production use.
And this part of the proposed definition makes it a blanket definition for all computer software not just AI: "Any artificial systems that perform tasks under varying and unpredictable circumstances, without significant human oversight".
So really hard to see how you regulate "AI" without a blanket regulation on all software development.
If we are talking about simulating complete multi-functional animal brains, especially human, then I think ethics do come into play. Perhaps our discussion should focus on that as something that should be regulated.
I think we have an societal interest in working to prevent the abuse of animals and people. And it could be that at some point, maybe very soon, we can effectively simulate a human or large animal brain and even good people might fail to realize the real perception of suffering, real suffering, they are causing in a thinking being stuffed into a computer.
That said, do we really want regulations preventing AI from becoming more like us? Is this inherently wrong? As every parent is acutely aware suffering is part of life and learning and we feel for our children because we have been there and understand how hard it can be. It is hard to imagine the human brain learning without negative feedback, without at least some bare minimum of physical and emotional pain.
Is the greater good in preventing any suffering or just limiting it to what is absolutely necessary for us to learn? It seems preventing all suffering is no different than preventing life. And allowing suffering more than what is necessary for life is also wrong.
Is there a golden mean between these extremes? And can that be regulated through the force of government?
You will know it when you have it as you won’t be able to regulate it.
The universe creates people to investigate itself. The universe is infinite and people are the smallest things available.
How can AI help the government? Well, if the government tries to regulate it then it won't help. However, if we replaced government with AI, with a system that actually learns, doesn't mistreat women, has restraint and doesn't bow to every lobbyist that shows up with a cart full of money, there may be hope for humans. But someone would have to program a system like that....nope, we're fucked.
Just as certainly as there are varying levels of natural intelligence, there can be varying levels of artificial intelligence.
Now if you want me to define "intelligence".... well, there's a trickier one. Is little Billy intelligent because he learned how to to multiply, or was that just the result of memorization? Is Alphazero intelligent because it learned how to play its games very well, or is it merely the result of following heuristic algorithms that coincidentally create the sufficiently persistent illusion of being a superior games player, while in fact possessing absolutely no real skill?
The answer is subjective... its going to depend on who you ask. Personally, I think that both are examples of intelligence, and more generally, any sufficiently persistent illusion of the existence of a thing, by virtue of being indistinguishable from that thing, should be considered completely equivalent to that thing, or else whatever we happen to call that thing doesn't really mean anything in the first place.
File under 'M' for 'Manic ranting'
Paperclips... Lots and lots of paperclips.
https://www.theverge.com/tldr/...
Nouvelles de jeux et technologies en français. TC
Well I was going to code it to figure out when the player plays which slots they have more of a miss on, So to when there was a tie, it would pick something that the person may have a better memory with.
If something is so important that you feel the need to post it on the internet... It probably isn't that important.
A system that has the capacity to say "I don't know the answer to that, but let me learn some more and I'll get back to you."
AlphaZero is a kick-ass Chess/Go player, but show it a pic of a hummingbird and say "what is this?" and chances are it's going to return an error state.
In other words, a program/system that can re-program its own code, or improve itself without external forces acting upon it to do so.
So rise up, all ye lost ones, as one, we'll claw the clouds.
Politicians are not even artificially intelligent?
Custom electronics and digital signage for your business: www.evcircuits.com
IMHO, self-aware just means being able to examine some internal states and store or report that information.
"Thermostat, what's the current temperature?" :: "72 degrees" - that's self-aware. It just isn't much of a self.
Even adding "able to modify internal states based on examining them" is something more than self-awareness.
Self-aware is not the same as "conscious." Consciousness implies assigning internally conceived meaning to, and abstract manipulation of, such states.
I've fallen off your lawn, and I can't get up.
Computer systems will become more and more capable, and will be entrusted with more and more tasks, ultimately being able to do just about anything we would like or need them to do. And there will be arguments over whether they are truly "intelligent", because all these things will be done via algorithms which are well-understood, from a foundation designed by humans. But it really will not matter if it passes someone's definition of "intelligence" at all.
-----
Sorry, I'm only a 1336 h4x0r.
This will only lead to a lot of people learning to speak with a Scottish accent.
I've fallen off your lawn, and I can't get up.
we should just ask Alexa.
I can't, I'll have to get my kids or my wife to ask Alexa. Alexa still doesn't recognise a British accent (at least not in the US). If I ask Alexa what AI is, I'll probably get a pizza delivered.
That's because Alexa only speaks English.
Just another day in Paradise
What does intelligence actually mean? ...and why is it so scarce?
It's all in your head. But not you specifically.
Just another day in Paradise
Ultimately AI is just glorified statistics or statistics that are de-glorified to make processing practical, as in lossy-but-fast. The field is new enough that it's hard to know what the edge cases will be in the future. "Think like humans" smells too fuzzy to me such that it would probably come down to opinions of the jury and/or judges.
As an example of fragile laws, you think it would be easy to write misuse of classified info (secrets) into law, but as the "Hillary email" case showed, it's far from trivial. There was the "intent" side of the law, but judging intent is subjective. This would also apply to "intent to create AI". (No significant evidence of H's intent was found in my opinion.)
And then there's the "gross negligence" angle that was analyzed, which is also subjective. For whatever reason, H was not given a complete security course, only a "briefing", such that she wouldn't be expected to recognize security markings (which resemble other legal documents, by the way). But is it "gross negligence" to not make sure she found a way to attend the course(s)?
In most orgs I know, assistants are expected to manage such schedules, pre-fill server-request forms, and remind the head hanchos as needed. But somebody not exposed to that may say the burden of everything is ultimately on the Chief because they are at the top. I agree she was negligent, but "gross" is a stretch. No one person can micromanage the requirements of the entire policy manual. That's just beyond the human head (except perhaps rare individuals with eidetic memories).
I've asked others how they would fix the secrecy laws, but so far every suggestion created new problems.
My point is that laws about tech and info end up being too subjective, and ultimately judged by personal experience in the work-place and elsewhere.
Table-ized A.I.
... with a PRETEND solution.
I think the DNC should run someone who puts this at the top of their agenda.
Ordinary people in that party are used to having their daily concerns overlooked.
Define "intelligence" first.
Before you can do it on a computer, it's A.I.
After you can do it on a computer, it's no longer A.I.
That's been true the last 25 years. I don't see why it should change in the future.
She was like chocolate when she drank... semi-sweet at first and then increasingly bitter.
That war was fought and lost a long time ago. It was found that I type development was too dangerous for the common people, and legislated out of existence. Just think about how you were taught 'to think', then take a look at the people around you, and ask yourself, is any of this really necessary?
A set of instructions which improves upon itself and repeats.
The important thing is "what entity is liable" - example, self driving car wrecks, chemical plant explodes, some "AI" messes up and does damage - who is responsible? Would it be the guy in the driver's seat of the AI car who was instructed to pay attention anyway and resume control if something wasn't right? The maker of the car? The guy who set it to "auto", assuming there was a choice? The guy who made a sensor that failed? The outfit that wrote the software? The outfit that promised that it would work fine? We missed the boat on software generally - Bill Gates would be in Gitmo at best for the man-lives he's wasted by proxy - because as mentioned above, by the time the issue arose, there was too much money in it, and software gets away with zero liability even when the author was extremely negligent. Of course, it's not that simple or no one would write or release any code, but the current zero, wild west setup ain't so hot either.
Why guess when you can know? Measure!
I have done some work with neural networks to operate as a surrogate model for more complex simulations. There is no analytical solution but some function exists such that for the given inputs the simulated outputs are the result to some determined precision. Neural networks are quite good at these kinds of problems. The system is in no way intelligent and it is not capable of doing anything beyond giving the same output as the simulation based on a given input.
I don't see any reason that just having a neural network involved should involve regulation.
Computer modeling for biotech drug manufacturing is HARD!
Interesting perspective...
You're claiming to define the limits of a prospective technology you've never laid hands upon, because it does not yet exist.
This is precisely like a caveman trying to explain the nature of the telephone.
I've fallen off your lawn, and I can't get up.
If we are concerned that AI will be used to mistreat actual living human people; then maybe Government should pass laws dictating the proper treatment for actual living human people. Rather than try to make these abstract definitions about the metaphysical properties of toasters and how they must be manufactured to behave.
The fact is - we've already wrestled with this problem. Our most primitive AI; the landmine. Kills or maims people. Rather at random. We tried to ban them worldwide. That effort failed spectacularly. There are apparently higher priorities in government than making sure random farmboys don't get limbs blown off when working land that was mined in a conflict 20 years before he was born. Dumb landmines. Dumber policymakers.
These are my friends, See how they glisten. See this one shine, how he smiles in the light.