Re: designing socially-constructive spaces

[David Galiel <web@galiel.com>]

It seems to me that the "uncanny valley" applies not only when talking 
about digital avatars, but also when talking about social constructs and 
interactivity.

It is far easier, and more reasonable, to use procedurally generated 
visuals and behaviors when dealing with unfamiliar contexts. You can 
feasibly convince a viewer that your Centauri mind-slug looks and sounds 
as real as its imagined surroundings.

So, I can see how Spore could make clever and productive use of 
procedural content at certain phases, but note how there is still, at 
the meta-level, built-in mechanisms for humans to share designs and 
creations, (and hopefully to tell human stories).

Far more difficult and perhaps even unproductive to focus too much 
effort on trying to recreate the most human-like 
behaviors--story-telling among them. Particularly when, embedded deep in 
every human is the instinctive desire to tell and be moved by human stories.

Artificial illusions of intelligent dialog, emotion, and particularly 
creativity are not convincing, particularly not when one can easily 
compare them to the real thing. Being "almost" real is often worse that 
being clearly machine, because of that uncanny valley phenomenon.

That is why NPCs that appear superficially like other players are a deep 
design error, in my opinion, and why, in the worlds I am designing, 
everything that looks human, is human, and everything that is AI/AL 
driven, looks like a machine or a non-sentient life-form (and, in fact, 
anything that is supposed to be an AI is, in fact, a human-operated 
puppet). It believe this avoids disappointing people - especially in a 
culture that, due to Hollywood science fiction, expects true, sentient 
AI *today*. I have yet to hear a compelling reason why we have to devote 
so much time, energy and budget to making artificial humans, when we 
have a potential market of hundreds of millions of real humans with a 
multimedia PC and an Internet connection, just dying for an opportunity 
to get involved in network-mediated collaborate storytelling--but 
uninterested in the current crop of "interactive" entertainment.

Even when talking about procedurally-generated visuals, such artifices 
should be employed, in my understanding, only when the human-generated 
alternative is highly impractical. It makes sense for Spore's terrain, 
less so for a virtual art gallery. We should not be so hasty to automate 
for automation's sake.

- galiel