Now there’s a ‘there’ there! Artificial imagination has arrived.
This was written largely — and fittingly — by Copy.ai and its machine-powered brain, in response to my prompts.
Introduction
I’ve always wanted to live in a town named after a food. I love the idea of living in an idyllic town where people are sweet, unassuming, and prone to bouts of contentedness brought on by the feeling that they inhabit an entire town named after a delicious thing. Sadly, there aren’t many towns like that: we have Boston (after beans), Hershey (after chocolate), and Fall River (after potatoes). But now — with the advent of artificial intelligence — we can invent as many places with delicious names as we want!
What is dall•e 2?
Dall•e 2 is a machine learning system that can create new images from textual descriptions. It can generate and imagine things in response to what it’s told. It’s one of the first AI attempts to create, to have thoughts, without a care about how those thoughts may or may not resemble the external world.
Dall•e 2 is an artist without prejudice — a potential source of machine consciousness.
What is artificial consciousness?
The ability to be aware of self is a simplistic, last-century notion of “object awareness” that might not even be how we have self-awareness. And as we get closer to this new foundation, we’ll need to recognize that what it means to have access or understanding of something isn’t just binary. It’s not just yes or no — you either know or you don’t know.
There are many levels of understanding and access — and the more you understand something (and the deeper your level of understanding), the more access you have to it. A simple example: if someone asks you about a movie that you’ve seen but haven’t watched in years and can barely remember, your level of access would be low; however if someone asks about an experience from childhood that has stayed with you for decades and influenced many aspects of who you are now (like my love for science fiction), then your level of access would be high because the experience has been influential across multiple areas of life rather than just one situation at one time.
How did this happen?
How did this happen? It’s actually pretty simple: the team used a technique called generative adversarial networks (GANs). Using these powerful tools, they took in millions of images and captions, while the system created a multidimensional space that humans can’t even fathom.
Then, using its multidimensional space to generate new things, new thoughts — including those we haven’t seen before — the system starts creating images based on understanding categories not only from previous examples but also from associations with other concepts as well. For instance, they’re able to figure out that if you ask it to “draw a tree,” sometimes what it will draw is not just any tree but one that looks like something you would see in an illustration for children’s storybook or cartoon show about forest animals!
What is artificial imagination
Artificial imagination is the ability to make connections and see relationships to create something new. It’s the starting point of consciousness, not the result.
In this sense, it’s not so different from creativity or genius — the ability to see things differently from others.
But there is a difference: In artificial imagination, your mind has been enhanced with technology that can help you find or create these new ideas faster than would be possible otherwise.
It’s hard for us to imagine what this will mean for culture and society as we move forward in time — just as photography was hard for people alive at its advent (and those who came before). But it does signal an important change in human culture: We now have a new tool, a new creative partner wielding which we can invent our future destiny together.
What do we hope comes next?
We hope to see a shift toward imagination. It’s not that we should give up on testing consciousness; it’s just that these tests have been marred by our own limited understanding of the complexity of human behavior.
Looking for consciousness among the behaviors of physical interfaces, like faces, spoken language, and bodies is like looking for an airplane among a trail of riven clouds: We know what it looks like when we see one, but we can’t recognize one unless it flies into view.
If we look instead at how people behave in spaces where they’re given total freedom to live out their imaginations — such as theater performances or films — we’ll get closer to seeing what real imagination looks like.
Conclusion
In the meantime, we’ll just be over here listening to our favorite Madonna playlist and waiting for our adventure to begin.
Written in collaboration with an artificial intelligence. You can try it, too, at https://www.copy.ai. Copyright © 2022 by Paul Henry Smith — whatever that even means now.
