I love when a classic science fiction concept becomes a practical reference point in current technology, and that happened during the most recent Town Hall in Second Life, when a user challenged Linden Lab CTO Philip Rosedale over his excitement around A.I., most recently seen with the company's partnership launch of AI-powered NPCs in SL.
Asking: "Look, I understand that Philip loves his A.I. and thinks it's the future, but I don't understand the insane push and force it in Second Life. Will AI rally around a creator when their pet is sick? Will AI support people when their parent dies? What do you have to say to that?"
Here's Philip's answer in full, replete with that perfect Blade Runner/Do Androids Dream of Electric Sheep reference:
I love that question, and how it's phrased.
I think, as I touched on earlier, A.I. means an increasingly large set of things I mentioned earlier. For example, that A.I. is used to do language translation. I suspect that most people here would like us to use A.I. for language translation.
On the other hand, as you say, A.I. avatars that are living in the world, brings up a very different set of considerations. Chief among those is, in the extent that Second Life is open and is generally designed to be a system where people can make their own /choices, for example, about content creation: How does A.I. fit into that?
And I think in that regard, we'll have a separate Town Hall, and we'll talk about it. What if we didn't want A.I. in the [virtual] world, what would that mean? Do we all have to take a Voight-Kampff test every few days?
But I welcome the opportunity to talk about that more, and I definitely respect the complexity of it. There's a lot to it.
I think the other [claim to] mention here is A.I. exists on stolen data. That argument, of course, is one that's going on in the world, big time right now. Like to what extent does the common crawl, which is the Internet as we know it today, to what extent should that be used for training A.I.s?
I again appreciate the complexity of that debate, and I look forward to the world stage answering it. You know, perhaps country by country, as we as we figure out what everybody feels, I think it's a fair and appropriate discussion.
I think the separate issues in a virtual world of why, or whether you use A.I. say to be a person is extraordinarily interesting and perhaps a conversation that's more uniquely ours to have as you know, the preeminent virtual world with adults in it, doing all kinds of different things, and building all kinds of different things.
So I think, I hope that [Second Life] can actually be a town hall and a forum in which good conversation and progress can be made about A.I., as I think we've been in many different ways over the years.
I hope to explore those topics with Philip further, especially because his answer here, while interesting, somewhat talks past the original question's intent. For one thing, AI-powered NPCs are still obvious, intrusive, usually annoying, and not anything like Rachael from Blade Runner. And I wouldn't plan on them getting massively better any time soon.
Even more key: Creators in Second Life are increasingly worried about the use of Gen AI in the content economy, while Gen AI-fueled "fast fashion" will probably erode the community/creator relationship that exists now with beloved virtual brands. That's the key topic Linden Lab needs addressing.
But what, then, is the right answer? I originally considered posing this as a reader survey, but gave up when I realized there were not many viable, non-controversial ways of regulating a technology that's already flooded the ecosystem.
More soon. Until then, stay tuned for that Town Hall which is sure to be a banger.
Please support posts like these by buying Making a Metaverse That Matters and joining my Patreon!
Wagner James Au / New World Notes 
Leave a comment