When David Usher – the Juno Award-winning musician, bestselling author and founder of Reimagine AI – visited Lethbridge Polytechnic last spring, his presentation left a lasting impression. It wasn’t just that he shared compelling insights into the evolving relationship between humanity and artificial intelligence; it was also the clarity with which he spoke about the opportunities and challenges ahead. His fireside chat on Artificial Intelligence and the Power of Creative Thinking was the latest installment in the polytechnic’s speaker series, presented by Bruce McKillop.
Usher spoke with the Wider Horizons team by phone a few months later, from his home base in Toronto, to dig deeper into the rise of AI, the role of creativity in a machine-driven world, and what it really means to leave behind a digital legacy.
When or where did your interest in AI begin?
I guess around 12, 13, years ago. I was doing a lot of keynote speaking, and I was talking to companies about creativity, but a lot about innovation as well. I really felt the next big thing coming down – the next big innovation or disruption – was going to be artificial intelligence. I was sort of onto it very early, and I don’t like to speak about things unless I have personal experience with them, like hands-on experience with whatever I’m talking about. So, I started looking for a project in artificial intelligence and I got involved with Google, working on a project to build artificial intelligence agents that could collaborate with humans writing song lyrics. And from that experience working with Google, I got the bug basically. And right after that, I started Reimagine AI.
As a talented songwriter, what was the process of working with Google Brain like? How did the songs compare to something you would have written on your own?
Well, this was before the real language model revolution of the last couple of years … so 10 years ago. The language models were terrible, and Google was very, very controlling about how they collected data. You really weren’t allowed to scrape the web like OpenAI has done and like everyone else has done now, and Google has done as well now. Back then you had to be very, very strict with your data collection. So experimentally, it was really interesting, but as a real tool that people could use, it was sort of meh. But then when OpenAI really scraped the entire web, that changed the ability of these language models to find real purpose and usefulness, utility.
Would you consider a song written by AI to be art?
I’m not personally interested in songs written by AI, just like I’m not particularly interested in art made by AI or writing made by AI. But that’s just a personal thing. Some people are really into that and those kinds of use cases of artificial intelligence. I use it for different things; in our company we’re using it for different things, but we’re not focused on that at all.
During your visit to Lethbridge Polytechnic, you predicted there would be significant disruption because of AI in the very near future. What steps can or should be taken to ensure the disruption is more positive than harmful?
We’re already seeing the disruption, really, aren’t we? The rise of agentic AI, which are sort of multiple agent systems that can work together autonomously; it’s really changing the nature of work, and how fast we work and how many people you need to do a job now. So we’re going to see a lot of disruption in the job market. We’re also seeing some very odd things going on with learning in relation to a lot of kids who are using AI now. We’re sort of in this nebulous place where we’ve just gotten through the big tech companies unleashing social media on kids without any parental control or government control, and now we’re unleashing AI on the same kids without any understanding of how it’s really going to affect the process of learning. Like, how do we learn things when we don’t need to go through the steps of learning to get the answer? I think there needs to be a lot of thought put into how we’re going to use AI in education, or limit AI in education.
Our creativity is an important characteristic that sets us apart from AI. How should we be leveraging that innate human quality within the AI world?
Creativity is what I call the creative thinking methodology, which is a creative process that you can apply to any vertical to take yourself through a process to come to an output. In this world of AI where AI can do the core work, you still need to be able to run a team and have the vision for what you’re trying to build. And the vision has to be more than generic, especially these days. I just read an article where companies are being overwhelmed by AI resumes. Every resume now looks the same because they’re just spitting it out through AI – make it better, make it this, make it that. And so every resume kind of looks the same and looks pretty good, right? So the creatives are the ones who are going to be able to take the average and make it amazing, and that’s why the process is so important.
Do you think we need to regulate the tech industry to address public good, public safety and protecting consumers?
One hundred per cent, yeah. I mean, we need regulation. You know, this is the most powerful technology with the biggest possibility of disrupting everything, and we need government to be involved with it.
Should tech researchers/developers be required to take ethics training and run projects through ethics boards?
There are a lot of people doing research into ethics, for sure. And we need research into ethics. Any time you’re about to release something with unknown consequences onto the world, you should look at the social impact of those things and how we’re going to deal with those things.
A request made through ChatGPT consumes much more electricity than a Google search. How have you seen the tech industry navigate environmental impacts?
I don’t really think it’s happening right now. I think that it should. I think that all the big tech companies are just on a massive build; they’re building the computers as fast as they can and getting the product out as fast as they can. I’m sure they have teams working on energy consumption as well, but it is a huge issue for sure. Now the hope is, and I think we’re already starting to see it, that as the models get better, like the big reasoning models, the smaller models, or the older models are getting cheaper and faster and hopefully more efficient.
In your presentation, you shared that Reimagine AI is working on virtual “echoes” of real people – a digital imprint that lives forever. What is the inspiration behind this technology, and how do you foresee these echoes being used in the future?
About four or five years ago, my friend Lucy – who’s one of my best friends from university and this incredible artist – was diagnosed with stage four lung and brain cancer. We built, basically a digital version of Lucy’s mind, an AI version of Lucy’s mind. And we have a touring exhibit where there’s – if you can imagine – a very, very large screen where you can talk to, almost like a light installation that represents Lucy. And as you talk to Lucy, and Lucy talks back to you, around the walls, all around you, are projected videos from Lucy’s life that are connected to whatever she’s talking about. So if she’s talking about her childhood, those will be reflected on the walls. If she’s talking about her life and dance, that will be reflected on the walls. It’s almost like you’re looking into her memory. And then [this past spring], the exhibit was in New York City, and her oncologist, who’s at Columbia-Presbyterian Hospital, came to the exhibit and was talking to Lucy live in the exhibit. She suggested we put Lucy online as somebody that people can talk to. They don’t have to tiptoe around anything; they can ask anything they want, or their loved ones can. And so that’s online now under askLucyai.com and we’re just about to start a clinical trial with a hospital in Brooklyn around this idea of the ability for Lucy to be able to help people who are newly diagnosed with cancer, so that’s interesting. And that is part of this whole idea that everyone in the near future, through our [SecondEcho] application, hopefully will be able to build a digital version of their mind, their thoughts, feelings and memories – really their history.
And just for fun, we asked ChatGPT to generate the final question for you; if your digital echo could interact with people 100 years from now, what’s one message you would want it to convey?
I mean, I don’t really think about myself 100 years from now but what I would say is, with SecondEcho, we’re trying to capture meaning within chaos. What’s your personal story? What means something to you? You know, trying to pull out those moments that have meaning because we’re constantly inundated with everything all the time, overloading our senses. So how do we capture things that mean something to us, that we can then connect to again and save.
