AI in the Metaverse: Utopia or Dystopia? with Louis Rosenberg

Speaker 1
We see that groups of people making forecasts or making decisions can get 30 40% more accurate when they're working together as a swarm as this interactive system with AI watching their behaviors. If they just took a vote or a poll or a survey and an example, we did a study with with MIT where we looked at groups of traders financial traders and we had them predict the price of gold the price of oil the S and P 500. And then they would do it alone, or they would do it just by taking a group vote, or they would do it as a swarm and as a swarm there's a constraint they have to all be connected at the same time. So it's a real time system that they would log in, it would question would come up as the price of oil in the next seven days and then we had these groups do this every week for 20 consecutive weeks. And the groups that use that work together as a swarm were I think 30% more accurate in forecasting oil and gold in the S and P, then the traditional methods.

Speaker 2
And then the question about it immediately, which is, does the system become more accurate with more people added to the system.

Speaker 1
So we had that exact same question first which was how many people do you need and I originally assumed that you would need at least 50 people, maybe 100 people. We looked at swarms of bees swarms of bees, it's interesting because a bee colony has about 10,000 members.

Speaker 1
And just like advertising in today's world in flat media, this type of advertising the metaverse will be targeted. And so you'll be approached by an AI powered avatar that will be custom designed to be most persuasive to you the way it looks, its voice, its mannerisms, its clothing will be all be chosen based on your profile. And then you'll be in a conversation about a product or service or political propaganda. And it will be very persuasive. And that use of AI I think is very dangerous, especially because in the metaverse that AI will also have access to your facial features, your vocal inflections, your lines, blood pressure, heart rate. And so we could find ourselves in a virtual future where we're engaged by virtual spokespeople that are driven by AI that have access to our personal history that knows our likes, our tendencies, it knows how it's likely to be persuaded, it's reading our facial expressions, maybe even our heart rate, blood pressure, the whole point of virtual reality and augmented reality is to blur the boundaries between what's authentic and what's not.

Speaker 1
The thing that's really different about targeted content in the metaverse versus targeted content in social media is that in today's world, when an advertisement pops up, it's very clear it's an advertisement. It's a pop up ad it's a video in the metaverse advertising will become immersive, it will go from black content to immersive experiences. And so when advertisements and promotion and propaganda become experiences, it could become indistinguishable from real events, real activities in that world, an example that I could give of what I would call a promotionally altered experience. Imagine you're walking down the street in a virtual world, or maybe in the augmented world. And, and you see a park car you've never seen before and you notice it and then two people are standing outside the car having a conversation about it and they're talking to drivers telling their friend how amazing that that this new car is and you just, you walk past and you might just think that was an authentic experience, and it could subconsciously influence you and say, Oh, like people who live around here really think that this car has these great features and you might not realize. No, that was injected into your world as a targeted advertisement. Only you saw that experience.