Blog

Clive Thompson of Collision Detection links to his excellent new gaming column at Wired. Thompson argues that the commercial viability of Bandai’s Tamagotchi, Sony’s AIBO, and now Nintendogs suggests that the real market for artificial intelligence is not in scary superhuman servants but in AIs that are helpless without our care and attention:

In reality, when robots finally broke out into the mass market, it was the Furby and the Aibo. Not only did they serve zero useful purpose, they actually demanded we spend hours and hours nurturing them. If you didn’t pay attention to your Aibo, it’d wilt. That, [MIT Professor Sherry] Turkle suggests, is precisely the reason these robots have such emotional purchase. Over in Japan, nursing homes are issuing Aibos to the abandoned elderly, because people love to feel needed — and as it turns out, that’s the one thing that Aibo is genuinely ‘useful’ for: making you feel needed.

In Authenticity: Brands, Fakes, Spin and the Lust for Real Life (2003), David Boyle contends that relationships with fake pets can never be more than a mockery of caring for real animals:

These are highly sophisticated toys. The animals may be marketed as robot pets — they can ‘learn’ after all — but people aren’t really going to have relationships with robots as they might with their pets. ‘There’s a major upside to i-Cybie’s low price,’ said Fortune, giving the game away. ‘You won’t feel so bad when you tire of it and drop-kick it across the yard.’

The same was true of those Tamagochi [sic] toys that had to be nurtured by pressing buttons, or they would died. Japanese restaurants opened créches to look after the Tamagochis belonging to their executive customers during their meal; schools provided bereavement counselling to former owners. But then the schoolgirls — who were supposed to be learning nurturing skills from these machines — soon changed the rules of the game to compete over how quickly and ingeniously they could kill them.¹

Even (perhaps especially) virtual people are at risk of real humans’ corruption by absolute power. Some players of The Sims are less than benevolent towards their creations. One gamer explains: ‘I guess people actually play this game to make their little sims happy. I’ll admit that i [sic] did that for awhile, but to be honest, it just got boring. So of course I reverted to my typical gaming pattern of torturing innocents to death.’ Will artificial pets ever be interesting or cute enough to stave off the caprice of their owners? Will there ever be a Society for the Prevention of Cruelty to Virtual Animals? Leaving aside the potential debates about consciousness, souls, and rights, I guess there are at least two preconditions:

1. Just like a dog or a cat, AIs will need to have a relationship with the rest of the world apart from their owner. There should be no off switch and it should not be possible to just put the AI away.
2. AIs will need to be genuinely unique individuals whose consciousness (or emulation of consciousness) is the result of somewhat miraculous processes like artificial evolution and learning rather than automatons directly programmed by a human being.

At present, game designers frequently remind players that they are, after all, only playing a man-made game. For instance, Thompson describes his experience with his virtual pet:

When you’re playing with your Nintendog, every once in a while it will — like a puppy in a store window — lean forward and rest its paws on the screen, trying to lick your face. It’s a rather sly, postmodern moment: For an instant, it seems, the dogs acknowledge that they’re inside a game. Even they know they’re not real.

Until designers stop giving the game away like that, virtual pets will continue to elicit the same level of mercy as Microsoft’s Office Assistant, Rocky the Dog.

1. Boyle, Authenticity: Brands, Fakes, Spin and the Lust for Real Life (first pub. London: Flamingo, 2003; repr. London: Harper Perennial, 2004), p. 196. (Available from Powell’s Books or Amazon.co.uk.) [back]