Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!


Forgot your password?
AI Google

WSJ Overstates the Case Of the Testy A.I. 230

mbeckman writes: According to a WSJ article titled "Artificial Intelligence machine gets testy with programmer," a Google computer program using a database of movie scripts supposedly "lashed out" at a human researcher who was repeatedly asking it to explain morality. After several apparent attempts to politely fend off the researcher, the AI ends the conversation with "I'm not in the mood for a philosophical debate." This, says the WSJ, illustrates how Google scientists are "teaching computers to mimic some of the ways a human brain works."

As any AI researcher can tell you, this is utter nonsense. Humans have no idea how the human, or any other brain, works, so we can hardly teach a machine how brains work. At best, Google is programming (not teaching) a computer to mimic the conversation of humans under highly constrained circumstances. And the methods used have nothing to do with true cognition.

AI hype to the public has gotten progressively more strident in recent years, misleading lay people into believing researchers are much further along than they really are — by orders of magnitude. I'd love to see legitimate A.I. researchers condemn this kind of hucksterism.
This discussion has been archived. No new comments can be posted.

WSJ Overstates the Case Of the Testy A.I.

Comments Filter:
  • by Improv ( 2467 ) <pgunn01@gmail.com> on Sunday June 28, 2015 @02:12PM (#50007243) Homepage Journal

    I'm calling the poster here out as being full of shit. As someone who's done neuroscience research, the idea that "Humans have no idea how the human, or any other brain, works" is bollocks. We have a reasonably good idea on the large scale, and in certain areas (such as the visual cortex), that understanding is quite far along. There are frontiers to our knowledge, but human understanding of brains is well on its way. Poster needs to pick up some neuroscience textbooks and get clued.

    As a particular recommendation, I'd suggest Kolb and Whishaw's "Fundamentals of Human Neuropsychology"; it's an excellent textbook.

    • You seem like someone informed. Can you tell me where to look to find out that actual likely abilities of the program the article is about?
      • by Improv ( 2467 ) <pgunn01@gmail.com> on Sunday June 28, 2015 @02:25PM (#50007317) Homepage Journal

        The WSJ article links a paper from some researchers at Google:
        http://arxiv.org/pdf/1506.0586... [arxiv.org]
        The WSJ article isn't particularly good either; they misunderstand what's actually going on in the research, which seems to be about conversational modeling (a "weak AI" type of research, the "understanding" being very shallow). They point out a few applications of this kind of work though, and that seems pretty solid/useful. (It doesn't approach the goals of "strong AI", those being actually modeling semantics and deeper reasoning)

        • by paskie ( 539112 ) <pasky.ucw@cz> on Sunday June 28, 2015 @02:39PM (#50007383) Homepage

          (I work in this area of research.) You are right, the paper is about just a sequence-to-sequence transformation model that learns good replies for inputs but is not actually "understanding" what is going on.

          At the same time, we *are* making some headways in the "understanding" part as well, just not in this particular paper. Basically, we have ways to convert individual words to many-dimensional numerical vectors whose mathematical relations closely correspond to semantics of the words, and we are now working on building neural networks that build up such vectors even for larger pieces of text and use them for more advanced things. If anyone is interested, look up word2vec, "distributed representations" or "word embeddings" (or "compositional embeddings").

          If you already know what word2vec is, take a look at http://emnlp2014.org/tutorials... [emnlp2014.org]

        • Re: (Score:2, Insightful)

          by Anonymous Coward

          The WSJ does research with on thing in mind, and that is to support Rupert Murdoch's personal political agenda. If only it were useful as toilet paper, it would be useful for something.

          And yes, i was a subscriber at one time.

      • Re: (Score:3, Funny)

        You seem like someone informed. You don't belong here is more like it.
    • by Pulzar ( 81031 )

      What are the prerequisites for understanding that textbook? Would someone with an EE degree be able to get something out of it?

      It sounds like an interesting read, but I hope that I wouldn't need a strong background in biology or chemistry to understand it, as I have neither. :)

      • by Improv ( 2467 )

        It starts gentle. I don't know if you'd enjoy reading the whole thing, but you'd probably get a lot out of it anyway. Good textbooks are like that.

    • by mbeckman ( 645148 ) on Sunday June 28, 2015 @02:37PM (#50007373)

      We are in the "cargo cult" phase of neurological research. Our level of cognitive understanding is like that of the South Pacific islanders who made bamboo replicas of WWII airplanes and radios after the GIs left. The islanders said to themselves "We must be very close to reproducing these wonders, because our airplanes and radios looks so much those of the GIs. Now we just sit back and wait for the magic goods to come out of the airplanes and wise voices to come out of the radios."

      If you really don't know how little we understand about the brain, NY Times science writer James Gorman can explain it to you:

      http://www.nytimes.com/2014/11... [nytimes.com]
      • Re: (Score:3, Insightful)

        by gweihir ( 88907 )

        Right on the mark. I have been following AI research closely for about 25 years now, an there is nothing that could explain intelligence. Not even a theoretical model that could work withing the constraints of this physical universe.

        At the same time, we can observe intelligence. An here is a little thing conveniently glossed over by some AI researchers and almost all neuro-"scientists": We can only observe Intelligence in connection with consciousness. Any actual researcher would conclude that the two are a

    • Improv,

      I was under the impression that visual operations in the brain were not understood at all. While we have a fairly good mapping of the visual areas of the brain and where things happen, we do not understand how images are stored or how we recognize (compare) images.

      Could you educate us (assuming bachelors degree level education)? I'm very curious how this works and how we would implement it in a computer system.

      • by Improv ( 2467 )

        The textbook I recommended above goes into this in much more detail, but I'll try to give a brief intro.

        The currently dominant map for understanding brain structure is the Brodmann map ; it's largely anatomical (clusters of densely interlinked neurons with mappable connections to others. The visual cortex is composed of brodmann areas 17 (primary visual cortex, containing a more-or-less bitmapped visual field), 18 (secondary visual cortex), and 19 (Third visual cortex). The visual cortex is divided into two

        • Thanx for the reply - I'll take a look.

          I'm surprised at the "more-or-less bitmapped visual field" comment because I would have thought there was something more sophisticated there - ie how do we recognize a cube when it's at an angle?

          • by Improv ( 2467 )

            Later brain regions parse information out of V1; the visual cortex is a pipeline (that forks in places). There are some great papers about people using neuroimaging techniques to pull an image out of V1. I think some of them have made it onto Youtube.

        • by gweihir ( 88907 )

          That is not an explanation. No engineer or real scientist would ever accept that non-explanation as one. This just says "we see activity in these areas related to it", without any understanding of the nature of that activity. Imagine people would try to find out how a computer works by looking at heat distribution during different activities. Sure, you could find where the graphics card was and and where the storage, but that is about it. It would be completely without any understanding of what is actually

      • Understanding how humans store and recognize images primarily is not a barrier to AI. It's not memory or image recognition that's the hill to climb; The fundamental algorithmic/methodological challenges are thinking, along with conceptual storage, development and manipulation (these things incorporate memory use, but aren't a storage problem per se.) Hardware needs to be able to handle amounts of ram and long term, high speed storage that can serve as a practical basis for the rest as well. Right now, we'r

        • Let's simply do a change up on a few words and query as to how the brain stores concepts and can compare, contrast and combine them. .
          • by gweihir ( 88907 )

            We do not even know whether it is doing that. Just that most people are capable of doing it to some degree. That is different.

    • by epine ( 68316 )

      Pretty much my reaction, too.

      Humans have no idea how the human, or any other brain, works, so we can hardly teach a machine how brains work. At best, Google is programming (not teaching) a computer to mimic the conversation of humans under highly constrained circumstances. And the methods used have nothing to do with true cognition.

      We don't even know enough to make the assertions quoted above with any confidence. Where's the precise boundary between programming and learning anyway?

      The prudent AI researcher

    • Re: (Score:2, Insightful)

      by mbeckman ( 645148 )

      Kola and Wishaw's text discusses the brain from two organizational perspectives, anatomical and behavioral. The authors never undertake to explain how the brain functions to produce the behaviors they describe. We thus know some of what the brain does, but nothing about how it does it. And the authors admit as much. Nobody knows how memories are stored, how vision is processed, how decisions are made. Science doesn't even know for sure that these functions occur inside the brain at all. There is,
      • by Improv ( 2467 )

        Souls are a myth from prescientific times. There's no point in contending with such concepts - they're part of history and superstition. If you don't understand brains, that's sad but correctable. There's a lot of research that you could read up on.

        Or I guess you could keep tossing that "cargo cult" term around and stay ignorant of the last 60 years.

        • "Souls are a myth from prescientific times."

          So sez the scientist... Do you see the irony?

          But that does get at why we will never see AI from digital computers; machines full of levers and switches that simply execute programs. Your program may become so complex that it is unpredictable, but that doesn't make it intelligent.

          • by gweihir ( 88907 )

            He is suffering from fundamentalist physicalism, a common thing among US atheists. They do away with God and then throw out all other things going vaguely in that direction, when there is zero need to. Hence these people fall for exactly the thing they think they are opposing: The use physical reality as their only true god and deny that anything besides it can exist. They claim Science tells us so, when it does no such thing.

            As an atheist and a dualist, I have zero problems with the concept of a "soul" or

        • A true scientist would not rule out an external force that could be termed a soul if it could be tested and measured. It would not be supernatural in that case, but part of nature. The term supernatural is a man-made descriptor for any phenomenon outside our current knowledge. Until Marconi discovered radio waves, the idea of transmitting information at a distance was considered supernatural. In reality, that misconception was just ignorance.

          Genomics, like cognition, is another discipline that may have
        • by gweihir ( 88907 )

          And there you fail. Is consciousness also a primitive, superstitious concept? Because Physics gives us absolutely nothing on it.

          You are just a fundamentalist physicalist, which is a quasi-religion. As all religious fundamentalists, you cannot actually grasp available evidence wherever it does collide with your fundamentalist beliefs. And hence your inane "explanations" (which really explain nothing) result.

      • by gweihir ( 88907 )

        There is also consciousness whis is apparently intricately linked with intelligence. From Physics, there is rather strong indication that consciousness is not part of the physical universe. There is just no mechanism for it. At all. With intelligence, it gets more murky, but half a century of failed AI research seems to indicate that matter and energy as known are actually not suitable to implement intelligence. The only known computing mechanisms that could approach some of the things that (smart) human in

    • by Megol ( 3135005 ) on Sunday June 28, 2015 @03:20PM (#50007557)

      We have no actual understanding on several important parts of the working of a brain, we don't know how memory works, we don't understand how decisions are made (or even what it means if one want to get philosophical) and we don't understand how an intelligent being get the feeling of self.
      There are a lot of theories and clues of how some mechanisms work (parts of how some levels of memory works, parts how neurons and synapses work, part of where and how some functions of the brain works, and even some mechanisms of self awareness). But that doesn't mean we actually understand it as a brain.

      Mental problems and physical problems in the brain aren't really treatable at the moment. What is done is the medical equivalent of carpet bombing with drugs that have little (if any) experimental proof of helping, for some cases they help - for some not. Side effects can be serious in many ways.
        One of the most efficient and oldest treatments available is that of ECT (Electro Convulsion Treatment) which again is a carpet bombing equivalent that causes a (somewhat) controlled seizure in the brain. But even that is really done without a thorough understanding of the working mechanisms - what is known is that it is often successful for a variety of mental problems, that it works quickly compared to drugs and some details like that of signaling substances being released during the seizure and that neural growth is increased in some parts of the brain. But again understanding of a few pieces of a puzzle doesn't mean we can even begin to comprehend the puzzle as a whole. How does it work? Anybody that claims to know is a fraud.

    • Might you then explain to us how a neuron weighs the incoming signals and decides which axons to direct its outgoing signals to?

      I know you gave yourself a caveat with "large scale", but those large scales don't really cut it for understanding how the brain fundamentally works, and in order to replicate its functioning in code (develop actual artificial intelligence), we will need to understand that. Add in how the effects of the various chemical baths it is subjected to modify cellular functioning and t
      • by Improv ( 2467 )

        The WSJ article isn't very good (as I noted in another comment); my comment here was mostly that we should also dismiss the commentary that the slashdot poster put alongside it.

        We know what most regions of the brain do. We have the ability to record some parts of the brain (at various levels) and have models that can predict activation levels based on subtasks. In the visual cortex, there are even people who can decode significant bits of the signal in V1. This is significant knowledge. It's not vague, and

        • by gweihir ( 88907 )

          No, you do not know anything of that sort. You do know that there is observable activity in certain regions of the brain when people do certain things. That is completely different and you claim means you do not understand your chosen field. You are basically claiming to know that the web-browser is creating the WWW, when it merely is an interface to it. At the current level of scientific understanding it is not possible to make the determination how much the brain is an interface and how much it is actuall

    • by gweihir ( 88907 )

      The poster is right on the mark. Neurosciences keeps lying to people about their great discoveries to reap funding. In actual fact, they have no clue how anything intelligent the brain can do works. They have so little clue at this time, that they can still not even be sure it is the brain that does these things. Physics and Mathematics and AI research seem to indicate that the brain cannot actually be intelligent, far too small and slow.

    • we may have some ideas about how the brain works — at an electro-chemical level — it has been well studying and documented. a good text would be by neurologist — john eccles:

      http://www.amazon.ca/Evolution... [amazon.ca]
      http://home.earthlink.net/~joh... [earthlink.net]

      as for treating a simulation of the brain as having the same qualities as a real functioning brain is to fear getting wet from a simulation of a rainstorm. there are scientists which would disagree that human consciousness is actually simulable in this w

  • by Vokkyt ( 739289 ) on Sunday June 28, 2015 @02:17PM (#50007271)

    http://arxiv.org/pdf/1506.0586... [arxiv.org]

    The actual paper isn't about AI much at all as it is about making neural conversational models, basically, having the computer chat-back at you in a prompt and natural way. The conversations are less about the computer responding cognitively and more about responding human-like based on the speech patterns fed into it.

    The researchers tested two types of datasets, an IT Help Chat Scenario fed with data from what I'm guessing are chat databases, and the second set was fed with conversations from movies as found from OpenSubtitles dataset (not sure if this is a relation to open subtitles.org).

    The machine took this vocabulary and then pumped out conversations, and the researchers just looked to see how the new sorting method worked.

    I don't understand the linguistic terminology nor the modeling at all, but it seems to me that this is less about AI research and more about just getting bot to sound a lot more natural when they generate responses. I guess this eventually has AI implications, but the research paper itself never even mentions AI, nor does it seem like that's their focus. They're just working on speech, and the statements the machine regurgitated were tested not for cognizance or sentience but coherence. The machine spitting out something relatively snappy isn't the machine getting an attitude, it's the machine finding something relevant to the input that the reader takes as snappy. Such an event has no more significance than when people trained Cleverbot to respond to questions about Hitler with "Hitler did nothing wrong". This bot is no more snappy than Cleverbot is a neo-nazi.

    • by mykepredko ( 40154 ) on Sunday June 28, 2015 @02:46PM (#50007415) Homepage

      I would argue that the process we have gone through here is a demonstration true intelligence at work.

      The original reporter looked at the article, didn't understand a piece of it and asked an intern specializing in technology what this was about.

      The intern couldn't be bothered, saw that it was a computer responding to human input and said it was "Artificial Intelligence".

      The submitter read the article and keyed on the comment about this being a machine learning, which they feel is impossible.

      Most /.ers (me included) responded to the submision and railed on about the ignorance of the media and the great unwashed.

      One poster actually read TFA and pointed out that it has nothing to do with the article, submission and most comments.

      I don't know how the hell we expect to create software that follows a process like this.

      • My original post complained about the WSJ hyping Google's research (read the title). I read both Journal pieces and Google's published paper. I suspect Google is as much to blame for not correcting the Journal's misconceptions. But my overriding concern is that this AI inflation seems to be happening with more frequency, and the hype is getting exponentially more hyperbolic.
        • A lot of tech is inflated. What about nanotechnology. We are not 10-15 years away from working medicinal nanobots.
          Remind me to beat Michio Kaku since you can't get a custom grown organ yet, like He predicted 5 years ago.
      • The submitter read the article and keyed on the comment about this being a machine learning, which they feel is impossible.

        ..which seems odd since there is a growing belief that intelligence is an emergent property of a particular subset of learning mechanics, and this isnt so much because of the vast knowledge we "understand" about brains so much as it is about the limitations we know must exist. For instance learning in brains must be primarily accomplished by local operators since the connectivity in the brain is primary local, and there cannot be many such operators since the brain is composed of essentially only 4 kinds o

      • by gweihir ( 88907 )

        Well said. The thing is that "human intelligence" is usually not very good. It is just the best thing available by an extremely large margin.

    • by gweihir ( 88907 )

      Indeed. This is Eliza on steroids and interesting scientifically. It has nothing to do with intelligence or cognition though. It is about making machines more interactive in ways accessible to non-experts. The machines remain machines.

  • If the media can't accurately explain to people and have them accept where AI really is, they only have themselves to blame.

    People have watched, kind, funny, evil, enigmatic machines interact with their favourite characters for years and have been told that true AI is just five years away for 30 years now.

    They've read about things like putting a worm's brain in a Lego Mindstorms: http://www.sciencealert.com/wa... [sciencealert.com]

    So, why wouldn't lay people believe ridiculous statements like "teaching computers to mimic some

    • If the media can't accurately explain to people and have them accept where AI really is, they only have themselves to blame.

      But the media gets just about every technology and science wrong when it comes to accurate reporting. AI is no different, why expect a different result?

    • by Burz ( 138833 )

      Perhaps its the media's fault for providing such bad raw material for the program in the first place. They condition it to be a movie junkie (presumably with a short attention span) and then expect it gracefully handle a _philosophical_ discussion? They might as well have asked it the secret to world peace.

  • Never mind. I'm not in the mood for a philosophical debate.

    • Never mind. I'm not in the mood for a philosophical debate.

      But...you started it.....

      I think the media needs to stop oversimplifying science and technology to the point where the average joe thinks he really knows something about it, not realizing all of the devils in the details, assumptions make a difference, risk are overstated or misrepresented, etc..

      Reporting should to such a depth that that the knowledgeable gain more insight, and less knowledgeable people realize their limits of understanding and decide to put in the work to learn, or leave it to the qu

  • I'd love to see legitimate A.I. researchers condemn this kind of hucksterism.

    I'd like to see legitimate A.I.s condemn this kind of hucksterism, myself.

  • mbeckman fails to grasp the core concept behind machine learning and AI. They aren't programming a computer to do things, they are programming a computer to learn things (or at a more advanced level, are programming a computer to learn how to learn things).

    He dismisses the whole concept like it is some kind of mechanical turk, but it is real, and it is getting better every day.
    • (or at a more advanced level, are programming a computer to learn how to learn things)

      That is pure hope as well as circular. I don't think he's dismissing the concept so much as dismissing the current field operatives as being anywhere near as far along as they promote themselves to be. He's doing it with a heavy dollop of derision, but I'm seeing that from the opposite side as well. .

  • by wytcld ( 179112 ) on Sunday June 28, 2015 @03:02PM (#50007475) Homepage

    The same articles show up over and over. The first states that computers are about to do consciousness. The second states that consciousness is a mere illusion for humans, whose actions are truly run from deterministic unconscious processes. In both articles, there is some hero scientist, with the article most often based on that scientist's press release.

    There is never a popular press article about how computers may never do consciousness, at least by any current definition of "computer," nor an article about how there are things human consciousness can do which no deterministic process can more than imperfectly mimic. Both of these positions are viable, and embraced by experts in various fields. By all current evidence, they may prove right. But it doesn't make for a hero story to write about someone who argues for these positions. "Discovering" that consciousness either essentially does nothing or that some computer advance is just about to do consciousness (or both!) is a "great" story. Editors like it. The public is impressed by the "brilliant" "counter-intuitive" revelation.

    • there are things human consciousness can do which no deterministic process can more than imperfectly mimic.

      Like what? Serious question.

      • by gweihir ( 88907 )

        The thing is that the "illusion" explanation is completely bogus. Even neuro-science is not claiming that. What they claim is that "free will" is an illusion, but they are doing so without good evidence and likely with serious misinterpretation of the data they have. (And good CS researcher can come up with several alternate explanations for what they are seeing. These people are not engineers and barely qualify as scientists. They have real trouble modeling information processing and they jump to conclusio

    • There is never a popular press article about how computers may never do consciousness, at least by any current definition of "computer,"

      If you look up my previous posts here on AI, you'll note that I'm pretty critical of the kind of press given to AI as well. And I think that we're pretty far off from a model of computing that will effectively rival the kind of learning the brain does.

      But even I think your claim here is asking the wrong question. If "consciousness" can be created using machines, it will be an "emergent [wikipedia.org] phenomenon," which means the kind of complexity that will appear may be sudden and unpredictable compared to the lower-

    • If consciousness is mere illusion, who is the illusion fooling?

  • As someone who enjoys programming computers to play strategy games (I highly recommend the General Game Playing MooC at https://www.coursera.org/cours... [coursera.org] for anyone else interested in this hobby), I do concede artificial intelligence has a long way to go before it's a match for natural stupidity. But AI is not all BS.

    While I have no idea how Google's algorithms work, this does sound suspiciously similar to the old Emacs game Eliza (https://en.wikipedia.org/wiki/ELIZA) whose original programer Joseph Weizenb

  • "And the methods used have nothing to do with true cognition."

    That's a bold assumption. The methods used for voice and image recognition certainly have a great deal to do with true cognition. It's certainly feasible that Google is playing with a true learning system and trying to teach and grow it rather than just throwing together another chat bot with scripts and trickery. Which isn't to say they've succeeded but just because none of the engines built to date have attained adult human level intelligence d
    • Knowing exactly how our own cognition manifests isn't a prerequisite to true cognition, a digital system could be completely unique in how it works and achieve true cognition.

      Or we could even come up with a system that works the way ours works without even understanding that this is how our system works as well... and maybe apply that information and learn something about ourselves. I was hoping that sentence would be a lot more coherent, but I'm not going to edit it now. First espresso in a while.

  • by PopeRatzo ( 965947 ) on Sunday June 28, 2015 @04:00PM (#50007675) Journal

    I suppose it was inevitable. My sex robot is going to make me sleep on the couch.

    I may have to go back to doing things manually.

    • I suppose it was inevitable. My sex robot is going to make me sleep on the couch.

      I may have to go back to doing things manually.

      You mean, program your sex robot in bytecode?

  • Even Eliza would do this... Sometimes she just got a headache and could deal with one more human complaint.

  • Does any know what the meaning of the word "mimic" is?

    imitate (someone or their actions or words), typically in order to entertain or ridicule.

    What part of "mimic" necessitates deep knowledge of the inner workings of a system? I can mimic a dolphin (EEEEK EEEK EEEKK QED), but that doesn't mean I have a clue how dolphins work. I was just imitating a dolphin to entertain you. It seems to me that the poster simply doesn't understand what the word "mimic" means.

  • This seems like a new version of the Eliza program with more memory.

  • In the future, whenever anything bad happens, people will ascribe it to the actions of a rogue AI. This will be great for corporate and government plausible deniability because they could program the computer to do exactly what it did but they'll just say that AI is too powerful and too complex for it to be controllable by us mere humans and we just have to live with the occasional bad outcomes. The high-frequency trading industry already tries to slide by with this excuse saying their market manipulation

  • Should have stepped right into the Monty Python argument sketch dialog.

Q: How many IBM CPU's does it take to execute a job? A: Four; three to hold it down, and one to rip its head off.