Reassuring parables

« previous post | next post »

The most recent xkcd:

Mouseover title:

'At least humans are better at quietly amusing ourselves, oblivious to our pending obsolescence' thought the human, as a nearby Dell Inspiron contentedly displayed the same bouncing geometric shape screensaver it had been running for years.

Anyone who spends much time trying to get computers to perform skilled tasks can come up with plenty of other reassuring parables. (Of course, anyone who spends much time trying to get humans to perform skilled tasks has a few stories, too…)

More than 30 years ago, Steve Levinson and I wrote a popular article about speech recognition research, in which we predicted that it would be a long time before computers were able to participate easily and naturally in human communications. But we also warned that when the time arrived, there would be serious social consequences, not all of them positive ones.

And some people think that this has already been happening, e.g. David Rotman, "How Technology is Destroying Jobs", MIT Technology Review 6/12/2013:

Given his calm and reasoned academic demeanor, it is easy to miss just how provocative Erik Brynjolfsson’s contention really is. ­Brynjolfsson, a professor at the MIT Sloan School of Management, and his collaborator and coauthor Andrew McAfee have been arguing for the last year and a half that impressive advances in computer technology—from improved industrial robotics to automated translation services—are largely behind the sluggish employment growth of the last 10 to 15 years. Even more ominous for workers, the MIT academics foresee dismal prospects for many types of jobs as these powerful new technologies are increasingly adopted not only in manufacturing, clerical, and retail work but in professions such as law, financial services, education, and medicine.

Update — I just got this notice for a Music Department colloquium tomorrow:

Why do we want our computers to improvise?

George E. Lewis
Columbia University

The loosely constituted field of interactive music has drawn upon AI and practices of free improvisation in creating a new kind of music-making that includes machine subjectivities as central actors. These “creative machines” have been designed to stake out musical territory, assess and respond to conditions, and assert identities and positions–all aspects of improvisative interaction, both within and beyond the domain of music. Meanwhile, theorizing relations among people and interactive systems as microcosms of the social promotes system design and real-time interaction aimed at achieving nonhierarchical, collaborative, and conversational musical spaces. The resulting hybrid, cyborg sociality has forever altered both everyday sonic life and notions of subjectivity in high technological cultures, and this talk opens up a set of question-spaces regarding the nature and impact of these transformations.

I'll be in the Netherlands, so I won't be able to attend, but maybe one of my computers will tell me about it when I get back.



29 Comments

  1. Bob Lieblich said,

    September 11, 2013 @ 5:54 am

    When do you think a computer will be capable of coming up with "If it doesn't fit, you must acquit"?

  2. Maslanka said,

    September 11, 2013 @ 6:23 am

    I wonder what we will do with those billions of unemployable people that are not only a surplus to the New World's requirements, but actually a load on it. What do you do if the commodity to be downsized is — people?

  3. mark said,

    September 11, 2013 @ 6:32 am

    I get a "failed to load PDF document" on the Scientific American website. Mirror?

    [(myl) The Scientific American archive, along with the magazine, now belongs to Nature Publishing Group, which of course wants $$$ for access. When I can get a copy for myself (Penn's library doesn't seem to have a subscription, and my subscription to Nature isn't good enough, apparently), I'll post it on my website and add a link.]

  4. Victor Mair said,

    September 11, 2013 @ 6:46 am

    But somebody has to design, build, and repair the computers and robots, and somebody (some human) has to tell them what to do. Or will there come a time when those functions too are taken over by the computers and robots themselves, as was predicted in 1968 by Stanley Kubrick and Arthur C. Clarke (2001: A Space Odyssey) — HAL 9000 controls the entire spaceship Discovery One?

  5. EndlessWaves said,

    September 11, 2013 @ 7:01 am

    With the result that conflicting information requires human intervention to salvage the mission.

    This is a win-win situation in the long run. Either computers take over basic needs and society is restructured to make jobs less important, or they don't, the balance of jobs adjusts, people retrain and everything carries on as normal.

  6. Victor Mair said,

    September 11, 2013 @ 7:12 am

    @EndlessWaves

    As your moniker aptly suggests.

  7. Alon Lischinsky said,

    September 11, 2013 @ 8:04 am

    @Victor Mair:

    somebody (some human) has to tell them what to do

    In fact, in current industrial processes most regulatory control does not come directly from a user, but from a feedback-enabled automated control system. That's what makes automation possible.

    Admittedly, while contemporary SCADA systems are impressively flexible and adaptive, they are very far from Skynet-type autonomy.

  8. KathrynM said,

    September 11, 2013 @ 8:10 am

    And, um, isn't the concern that machines will make human workers redundant as old as the industrial revolution itself? Which is not to say that the concern is entirely invalid; just that it is an old rallying cry, and there is no reason to see the problem as suddenly becoming exponentially more serious.

    [(myl) Well, the thing about the industrial revolution (or the enclosure movement, or the colonization of the Americas, or whatever) is not that it produced the end of the world or endless evil, just that there were a few problems, some pretty serious. Same thing this time around.

    It would have been a mistake to insist that machine-made cloth could never replace the work of well-trained artisans. And it might be a similar mistake to assume that artisanal lawyering or doctoring can never be replaced.]

  9. Jonathan Mayhew said,

    September 11, 2013 @ 8:18 am

    I'd like to see a computer that could design a human good enough to beat it at chess.

    [(myl) Excellent premise for a story. There could be a prize involved: the Mayhew Prize.]

  10. Corey B said,

    September 11, 2013 @ 8:34 am

    My own view has always been that I'm not so worried about when the computers are going to become sentient. I'm more concerned about when humans will.

  11. Gene Callahan said,

    September 11, 2013 @ 8:37 am

    "And, um, isn't the concern that machines will make human workers redundant as old as the industrial revolution itself?"

    I bet when plows replaced sticks there were a lot of people worried that this would throw almost everyone out of work.

  12. Ted McClure said,

    September 11, 2013 @ 8:51 am

    Re mark above: Try your public library and state library web sites. You may get lucky and find your tax dollars have already paid for it.

  13. A Reader said,

    September 11, 2013 @ 10:51 am

    Gene, one of the usual explanations for why plow technology improved in the Middle Ages was actually the other way around: economic changes from the massive slave-based labor force of antiquity meant that there was more pressure to use technology to maximize the labor of individual workers.

  14. J.W. Brewer said,

    September 11, 2013 @ 11:04 am

    But without the transitional dislocations of the Industrial Revolution (and contemporaneous pushback), we wouldn't have the awesome etymology (ok, ok, one rival version of the etymology . . .) of "sabotage," which is still doing quite well as a lexical item despite changed circumstances. I doubt "twerk" will be that vigorous a few centuries from now.

  15. leoboiko said,

    September 11, 2013 @ 12:27 pm

    Right now humans do a lot of manual labor or repetitive tasks, but these are increasingly being automated, while the average wealth is ever higher and consistently increasing.

    If the jobs disappear because robots and computers take them up, won't the total wealth stay the same or (more likely) rise even more? So we don't actually need all those people to be working all the time, right? So we could just turn social-democracy up to 11. Living wages for everyone.

    This blogger thinks that this is already happening, only our sense of work ethics/identity doesn't allow us to admit it; so we talk of unemployment as a problem, while simultaneously pretending that the living wages are "food stamps" or "Medicaid" etc.

  16. D.O. said,

    September 11, 2013 @ 12:32 pm

    Now, if computers learn how to read and comment on blogs during the time normally allocated for gainful activities… Wait a minute, what's going on right now.

  17. Rubrick said,

    September 11, 2013 @ 2:27 pm

    Ultimately it will be found that the only thing computers can't master is writing jokes, and everyone laid off from their now-unnecessary jobs will become standup comedians. It'll be just like the '80s.

  18. Brett said,

    September 11, 2013 @ 2:33 pm

    @D.O.: xkcd number 810 is relevant to that point.

  19. marie-lucie said,

    September 11, 2013 @ 3:12 pm

    JWB: the awesome etymology (ok, ok, one rival version of the etymology . . .) of "sabotage,"

    As a French speaker I was not aware that the word had more than one etymology. Un sabot can only mean a 'wooden shoe', the former daily footwear of great numbers of European peasants and laborers. I am curious to learn about the alternative etymology.

  20. Jonathan Gress-Wright said,

    September 11, 2013 @ 3:24 pm

    Increasing welfare to compensate for fewer jobs seems like a bad idea to me. You'll get a healthier society if everyone has something to do.

  21. Nathan Myers said,

    September 11, 2013 @ 4:04 pm

    @leoboiko: In fact it is already happening, except that the displaced wealth is taken up by the owners of the machines, while the displaced workers take up fast-food counter jobs, if they can get them. Since the '70s, the top 1% has taken up effectively all the benefits of improved production efficiency. The rest of us are no better off, or less well off, than our parents were. This is a new thing.

  22. Asa said,

    September 11, 2013 @ 5:16 pm

    marie-lucie: The Online Etymology Dictionary says that there was a chain within French from "sabot" meaning "wooden shoe" to "saboter" meaning "to sabotage, bungle" (literally "to walk noisily") before the word even entered English. As far as I can tell, the Online Etymology Dictionary doesn't cite sources in specific entries, so make of all of this what you will, but it explicitly says that "the oft-repeated story (as old as the record of the word in English) that the modern meaning derives from strikers' supposed tactic of throwing shoes into machinery is not supported by the etymology."

    I'm a little disappointed — I quite liked the image of the French workers throwing their clogs into the machinery.

  23. Eric P Smith said,

    September 11, 2013 @ 5:32 pm

    Oxford Dictionaries Online gives French ‘saboter’ not as “walk noisily” but “kick with sabots, wilfully destroy”.

  24. a George said,

    September 11, 2013 @ 6:06 pm

    @Jonathan Mayhew:
    this reminds me of the old question regarding the Divine Being's omnipotency: – if He is omnipotent, can He then create a stone that is so heavy He cannot lift it? Possibly the short answer is "yes!"
    No prizes.

  25. Chris C. said,

    September 11, 2013 @ 6:33 pm

    @a George — The idea of a machine designing a human being is not the sort of logical contradiction that silly conundrum is trying to point up. We may see it as implausible — at present — but it may not necessarily always be so.

  26. Jonathan Mayhew said,

    September 11, 2013 @ 7:05 pm

    Right. It's not a logical contradiction at all. It is a practical "impossibility" but not impossible to imagine. My point is that the computer so sophisticated that it could create a human that could surpass it in some narrowly defined task like playing chess would be judged "superior" to its creation, but maybe inferior to the human who created the computer who created the human? It's not a logical contradiction for an engineer to build a plane that can fly faster than she can run, either. It has been done.

  27. marie-lucie said,

    September 11, 2013 @ 9:35 pm

    sabot, saboter, sabotage

    I should have consulted the TLFI (Trésor de la langue française informatisé) first. It appears that sabot has various other meanings, mostly having to do with thick, coarse or ungainly objects as well as specifically those wooden shoes, always made much larger than the intended wearer's feet in order to leave room for straw, hay or in modern times even slippers, used as insulation. The first meaning of the verb saboter is given as 'make noise while (or as if) walking in wooden shoes', indeed producing heavy clatter on a smooth hard surface such as flagstones or a wooden floor. Most other meanings have to do with unskilled, clumsy handling and even destruction of materials. So the strike of the Lyon silk weavers which is famous for their sabotage of the newly introduced machinery which they feared would destroy their livelihood may have popularized the word, and those workers may have indeed used some sabots for that purpose (although as indoor workers they probably did not wear the heavy wooden shoes), but that was not the first time that the verb saboter was used.

  28. dainichi said,

    September 11, 2013 @ 9:52 pm

    IMHO, the problem isn't welfare or the threat of a robot takeover, it's the increasing gap between those who control the robots and those who do not. Almost all technological progress has caused concentration of wealth, and it usually takes revolutions of some kind to reestablish equality.

  29. Chris C. said,

    September 11, 2013 @ 10:42 pm

    I must confess I had the old Terry Bisson short story, "They're Made out of Meat" in the back of my mind.

RSS feed for comments on this post