Turing Chicken

JBJ points to what turns out to be a Burger King ad called Subservient Chicken. It is a bizarre concept for an ad, some mix between one of those annoying Hamster Dance type memes, a Turing test, a webcam (of the “what are you wearing?” variety), and a videogame avatar who gets bored (Spyro or Sparks looking back to see why he’s not been told to move recently).

Commands I tried:
dance
jump up and down
do the chicken dance [did a different dance, but not quite the chicken dance]
drink tequila [he walked to the bar and downed a drink]
stop drop and roll
shake your booty [same as ‘shake your tail feathers’]
wave your left arm [did a cross between a ‘flex’ and a ‘wave’]
touch your toes
play dead

I was impressed by the “drink tequila” answer, so I tried altering the environment, something that would potentially have a lasting impact.

Turn off the lamp.

The chicken looked over towards the lamp and then tilted his head, as though he was asking “do you really want me to do that?” ‘Turn on the TV’ had the exact response – the pan of the chicken’s head allowed him to cover about 1/2 of the room and then he gave that little, noncommittal shrug. But I asked him to clap, and it turns out the lamp is connected to the Clapper ™, because he claps, the lights go off, and then they come back on. Hmm. Ok.

Move the chair.

The chicken suddenly leaps onto the chair, sitting on the armrest, his back to me. Is he trying to move it? I’m not sure, but it doesn’t look like it. The chicken responds to the command “sit in the chair” by walking by the chair, obviously considering it too small for his tail feathers, and deciding to sit on the couch instead.

What intrigues me is the difference between the textual Turing test, where some ‘ELIZA’ (a generic example) responds via text, and this visual one, where a single response might be taken to mean several different things. Drink tequila, for example, could be just “have a drink” or “drink a coke,” and the responding image could be exactly the same. Is it really different from “ELIZA saunters over to the bar, pours a drink, and gulps it down.”? You can’t ask the Subservient Chicken to describe the taste, but that’s not a limitation when you expect a textual response. If ELIZA drinks a shot of tequila and then describes it as “sweet,” we have an obvious disconnect – not simple disobedience. Do visual cues require (or offer) less specificity … or better yet, in what ways can images and text, respectively, be more or less specific and get away with presenting believable behavior?

I’m also intrigued by the use of the imperative – we have a series of reverse commands, both given and received. The submit button itself functions this way; we both submit a command and are commanded to submit. A minor point, but one I’m working through (regarding games, not submissive chickens) in the dissertation …

I do agree with JBJ – I’m not sure I feel more inclined to eat chicken from BK, even if I can have it my way, but the subservient chicken is fun to boss around, if just for a little while.