Original virtual-reality machines sought to immerse people in 3-D graphical spaces, usually with head-mounted displays or other devices that gave 60 percent of those who tried it vertigo or nausea. The creators forgot that human beings crave social interaction. Today's virtual worlds (avatars and multiplayer gaming) don't try to immerse people. Graphics matter less than the quality of communication between people.


Embodiment of a person online comes through dialog, with the scenery and dress taking second place. And there is nothing "virtual" about the "reality" of people's interactions and lives in these worlds, where they meet people and experience events that touch them as deeply as anything in "RL" (real life).


Creating the virtual me


Most people don't like what they look like in photographs, or even in mirrors. In my six years of avatar experience online, I've rarely seen a real likeness of a citizen.

Presenting a digitally distorted human face is dangerous territory. Nothing weirds out a person faster. We will accept cartoons, masks, abstractions, even just straight old flat photos, but try to create a deformable, 3-D head and that's where some seriously ancient social neurons kick in.

I wish the many ventures that try to do photorealistic avatars, bots, or agents well in these efforts. Just don't burn through too much of someone else's money before you do some focus groups. Or at least enter real living avatar spaces and ask citizens what they prefer and why.


Talking heads


Like photorealistic heads, talking-head agents may also have a big problem gaining acceptance. When it comes down to it, it is very, very hard to create a convincing, entertaining synthetic character. People hate the automated phone answering systems many "customer service" organizations use. We're bombarded daily with impersonal interfaces from companies and governments trying to save money with cost-saving, frustrating systems. Talking-head agents come dangerously close to triggering the "Damn it, give me a real person!" response.

If a talking head is meant to entertain us, then fine, we'll sit back and watch the theater. If, however, we're trying to search for something, find an answer, or just get our work done, then these automated personalities better come with a quick "turn off" switch. I'm sorry for being so harsh on well-meaning, hard-working technologists, but does anyone remember (or use) Microsoft Bob, Open Sesame, or Clippy from Microsoft Office?

Even younger minds engaged in multiplayer gaming treat "non-player characters" to the sword, rail gun, or other ignoble end. They will certainly have a "subzero tolerance" for attention-soaking, slow-speaking, lowest-common-denominator talking heads.


The future of artificial intelligence I see in the current fascination with talking heads a return of the old dream of "Grand AI," a dream powered by four generations of engineers who tried to create an artificial human with a fully developed synthetic mind. But we have only vague ideas of what consciousness is and how the brain works, and it's likely you can't code human intelligence in a bottle, separated from its total environment. Also, our best engineers aren't very good at creating software. Only in Hollywood will we experience the enticing (or horrifying) visions of Grand AI.

I think the future lies in what I call Tiny AI. Bots and agents -- micro-intelligent prostheses -- will augment reality. They'll extend the capacity and reach of the human mind and community communications. I think an ethical issue will arise in which people will demand to know when a Tiny AI is talking and when a real person is interacting in real time.

My particular vision is that a person will exist in cyberspace (which includes wireless devices) as an avatar embodying their real-time presence. They will be surrounded by a cloud of Tiny AIs that proxy them into many processes and aspects of virtual community.

You can read the entire in-depth interview on virtual worlds here

Return to Digital Space Documents