I'm in Seoul at the moment, on an "AI Tour" of Southeast Asia in Hong Kong, Zhuhai (China), Seoul and Taiwan. Having spent a couple of years trying to get academic staff up to speed with what's happening with the technology, I've come here with a slightly different message from the usual "look at what you can do with GenAI" stuff. The message is how do we think ahead of where the technology is "at", to thinking about where it (and we) are going. To put it simply, this technology is going to change the way we perceive the world. In some ways, this is what technology has always done (it's striking to think that Chinese society is now unthinkable without the mobile phone), but AI is going to present fundamental perceptual shifts to us, and this will have a huge impact on how we learn and coordinate ourselves in the world.
If we assume that the fundamental scientific breakthrough with AI has been made (although I think more discoveries are on the way which will bridge the biological/technological divide), then what is going to happen next is a predictable increase in speed and scale. In terms of speed, the things that we are accustomed to taking a few minutes or seconds like image and video generation will become almost real-time. Speed change is a fundamental change in the nature of the technology: images appearing as we speak will change the way we communicate. Video appearing as we speak will be even more profound. At some point we may even have images appearing as we think. I would have been sceptical about this stuff a few years ago, but it some of this stuff is already practicable, and the rest is coming into view. We simply aren't ready for what it will do to us, and to a large extent we are worrying about the wrong things.
Interactive video also will be soon with us - so not only will AI generate video from prompts, but will enable us to virtually interact with that video. I showed this demo of a game world that was generated by AI in Seoul. It's already pretty astonishing. Computer games are going to become increasingly important - potentially as a means of communication. It's making me think that my scepticism about VR was misplaced. I had based this on the fact that the content for VR is so time consuming to create. But with AI content generation will become as much of a non-issue as it has rapidly become with creating text.
I can't say what this will all mean. But I can say that this is likely to happen.
