Canon PowerShot G9
ƒ/4.8
44.4 mm
1/60
200

Just outside the third-floor window of the IMMI board room is a large nest of crows up in a eucalyptus tree. The adults appear to pair bond throughout chickraising, and this baby looks ready to fly the coop.

They reminded me of one of my favorite TED talks, by Joshua Klein, on the intelligence of crows

Skip to Minute 3:00 if you want to see something amazing.

If you make it to the end and hear about his peanut vending machine for crows, think about some of the legal questions it raises. He put it our in a field and left it alone for days. He came back later and it was full of coins, which the crows “cleaned up” from the streets.

Does he face any criminal risk as the builder of the machine?

If the crows start bringing coins from café tables (stolen money), does this change the equation?

If not, what if he switched the training to paper currency on table tops? (a bit of purposeful design snuck in…)

If the trainer has committed a crime, what legal precedent would that set for unsupervised evolutionary algorithms and iterative training sets? Think of robotics and AI… It could have a chilling effect on the primary path to complex systems development in the future…

Could emergence become a crime? Only if predictable in advance? Only if purposeful?

16 responses to “Crow’s Nest (ruminations)”

  1. Interesting question. I’d think that criminal or civil liability would require either intent or some clear measure of predictability (or else we’ll be prosecuting butterflies for causing hurricanes!). Might be analogous to how the owner of a pitbull is held responsible when the dog attacks a neighbor.

    In any case, Joshua Klein’s TED talk is also among my favorites, as it takes.on the conventional wisdom that intelligence is only found in beings that resemble us in some way.

    Richard Dawkins seems to like the story of Betty the Crow too, which he reprinted on his website:

    In a challenge to man’s sense of the uniqueness of his own intelligence, an ingenious crow called Betty has managed repeatedly to twist wire into a hook to lift food from a tube in a British laboratory. "We had to convince ourselves it was not a fluke, so we repeated the test 10 times and the animal did it in nine of those," said an excited Professor Alex Kacelnik, who led the experiment at Oxford University. Showing an extremely rare capacity for an animal to understand cause and effect and create a tool out of non-natural material, the female crow bent straight garden wire — a material she had only seen before on cage meshes — into a hook.

    The researchers were testing whether the birds were able to lift food out of a vertical tube using either a straight piece of wire or a hook. "The surprise came in trial number five when the male stole away the hook and flew to another part of the aviary," said Professor Kacelnik. He watched as Betty spontaneously bent a straight piece of wire and used it to retrieve the snack.

    She held it in her beak to lower into a vertical pipe from which she lifted" a small bucket with meat morsels inside. That display of what the team of three scientists call "tool-related cognitive capabilities" has challenged previous assumptions that primates like apes were the best after humans in problem solving intelligence. "We assume primates will be cleverer because they are closest to us," Kacelnik added in a phone interview from his laboratory in Oxford. "But this animal (Betty) seems to be on a par at least with any primates we have seen." In the tests, Betty consistently outsmarted her older male crow companion Abel — both from the Corvus Moneduloides species on the Pacific island of New Caledonia. Fortunately for male pride, however, scientists attributed Betty’s superior brainpower to her relative youth, not sex.

    The only time in 10 experiments when Betty did not make a hook out of the wire was when Abel managed to bring the food up with straight wire. On other occasions, he waited for Betty to bring out the food then stole it from her.

  2. Indigenous traditions have long known the interactive intelligence of crows/ ravens and human beings. In the Pacific Northwest, Raven is a trickster, teasing and taunting human endeavor, or like in Haida tradition, Ravens also stay close to us, they are here to protect human beings.
    ….in Miwok myths, ravens became people; they became the Miwok of the Sierras.

    Very special birds.

  3. First time I see a crow’s nest. I while ago I could capture how crows are "cleaning" Tokyo’s streets eating rats, it looks like Tokyo’s ecosystem is pretty balanced:

    [http://www.flickr.com/photos/torek/3362053476/]

    While capturing this picture, Joshua Klein’s TED talk came to mind also 🙂

    Last calculations from Tokyo’s government estimate that the population of crows is around 50.000 in the city.

  4. Abel sounds like a manager I know…..

    There is a legend that if the ravens ever leave the Tower of London, the monarchy, and the kingdom, will fall. Today, those Tower ravens are BIG suckers and fearless, very well fed and cared for (the ravens have the Beefeaters well trained). The Brits cheat a bit – their wings are clipped so that they can’t leave, even if they wanted to. (see Wikipedia-Tower of London)

  5. That TED talk is one of my favorites. It’s the one I show people who’ve never heard of TED.

  6. complexify: the intelligence of Abel is intriguing as well. As for the legal issues, I am a lay person, so I asked these questions of legal eagle Mark Lemley, and I have lost his reply, but the gist of it was that there was no clear legal precedent. One of the key questions is predictability as you mention. Joshua builds a machine and leaves it in a field. Wild crows that he has never met, and which are not his pets (no property or control concepts), end up performing novel acts. Was this provably predictable before he built it? He thinks they are picking up litter. If they were actually "stealing" money from cafe tables, and Joshua did not know that, is he innocent?

    IMG_6706 IMG_6709
    IMG_6707

    restlessthought: I have a very cool talking stick carved by a native out on the Vancouver islands. It is covered with bird heads, and I wonder how many are crows vs. eagles?

    Oh, and I think they are technically ravens, the smartest of the crow family…

  7. pegleg000: I think we all know at least one or two Abels. 🙂 Abel does seem intelligent in his own way, though we may be anthropomorphizing. Unlike Betty, couldn’t Abel’s actions simply be persistence and then opportunistically stealing whatever food was in front of him?

    jurvetson: proving predictability would indeed be tough (assuming no "Quoth the raven" moments). Thanks for sharing Mark Lemley’s opinion. As a lay person myself, it’s good to know that even the preeminent legal scholars wrestle with these issues. I’d hate to see research into emergent behavior slowed by legal concerns, since by definition an experimenter does not have certainty on what the outcome will be.

  8. I teach squirrels to take cookies.

    I guess the supposed-trainer’s intent plays a big role.

  9. Hi ,nice pictures, seems like they can actually learn to speak a bit and say
    hallo whats up? 🙂

  10. … "Just outside the third-floor window of the IMMI board room " … really smart birds, figured out a good way to get great media exposure 😉

  11. yes… great out-of-home viewing.

    By total coincidence, as I was posting this, a new PNAS paper (by Christopher Bird no less) explored tool making in the lab…. as summarized in the Economist:

    they have discovered that rooks, which have never been seen to use tools in the wild, can learn to use them, and even make them, in the laboratory.

    The upshot is that toolmaking, at least in crows, does not look like a specifically evolved ability but rather an extension of general intelligence. Perhaps wild rooks are not presented with a need to use tools, and so don’t bother. What this implies for the evolution of human toolmaking is unclear. But it puts a new spin on the phrase “bird-brained”.

  12. and they recently found that they can do three sequential steps in tool making.

    Also, some Avian Einstein videos… =)

  13. sharp, great videos, thanks. So some bird is in a way sharper then a chimpanzee, and bird brain relate to human.

  14. "Rooks smarter than a 6-month-old baby"
    New Scientist headline, Oct. 10, 2009

  15. smarter than a 6-month old? how in the world could they ever control for limitations in body coordination and capability? a 6-month old can’t move around like a rook, can’t successfully manifest learning or cognitive objectives in the same way… also seems that the rook and baby would attend to and (innately) prioritize different stimuli. ..is interesting.

    also, i’ve always wondered about our moral standing among animals that are proven smarter than we are (even if only at various developmental stages)….seems especially strange to kill, or eat, a being that is smarter than oneself.

  16. what lovely blue eyes! thanks so much for sharing the link to that talk by joshua klein. it is just amazing what he has discovered. i first learned about crow intelligence from this nature of things documentary. i felt so guilty for ever yelling out my bedroom window for crows to be quiet in the morning. i think the idea of training crows to pick up litter for peanuts is the greatest and most possible idea i’ve heard in a long time. i have been feeding some in my backyard for a month or so and watching them so carefully stash their food bits and return for more is fascinating enough in itself. i still have lots to learn, but it is an enjoyable journey. all the best.

Leave a Reply to jurvetson Cancel reply

Your email address will not be published. Required fields are marked *