If you've been on social media recently, you've probably seen an AI-generated cartoon of someone you know. Creating images with generative AI is a hot trend, but one man who got into AI in the 1980s said that wasn't the original intent behind artificial intelligence.
Dean Allemang, of Columbus, is a computer scientist who currently works at ServiceNow, a cloud computing company.
Allemang was in his 20s and studying mathematics in Cambridge, England, when he decided to pivot and study artificial intelligence, which was considered the "hot new thing." He earned a PhD in AI from The Ohio State University in 1990. After graduating, Allemang became disillusioned with artificial intelligence. Now he's back to being an AI enthusiast.
Allemang talked with WOSU's Allie Vugrincic about the way AI has evolved and what role he thinks it plays in the world today.
Vugrincic: So what were your expectations when you first got into AI, and did AI meet that or not?
Allemang: I really wanted to understand what is it that intelligence is and how do we do it and understand that well enough to help human beings be intelligent better. And that's what was really driving me and I think the whole field in those days. And, to a large extent, no, AI did not deliver on that. But this is where I became really disillusioned. In 1995, something did do that. It really did help every doctor do their job better, every mechanic do their jobs better. Every intelligent person did their job because of an invention called the world wide web. The thing that made us interested in AI was how could you have the knowledge of the whole history of the world, literally at your fingertips? And the web did that in ways that was very different from what we were talking about in artificial intelligence in those days.
Vugrincic: So if the web did what artificial intelligence was trying to achieve, what purpose does AI serve today?
Allemang: I really feel that modern generative AI has really tapped into something brand new that we did not understand as little as 10 years ago. It all has to do with language. Language models are the things that suddenly are... people are talking about artificial general intelligence and they're talking about it seriously. No one's talked about that seriously outside of science fiction forever. But what's happening in terms of changing how we live is the whole thing of how you work with a computer. Back in the days of the web, yeah, we're driving it, we look things up, we use them. Now the AI is collaborating in ways that I did not expect to see during my lifetime.
Vugrincic: One of the pushbacks against artificial intelligence, especially using it in these social media trends to generate images or videos, is that it has a big environmental impact. AI uses a lot of electricity and a lot water, but you say it's not as bad as people make it out to be.
Allemang: I set up an experiment right here in my home. The software is open source. You can buy a GPU. And I set it up (to answer) how much does it cost to make an image? And I found that if I really run my rig as hard as I can, it's about the power output of an old fashioned 150-watt bulb. You know, the kinds of bulbs that you might have had in your basement back when you were a kid. So, you also have to put it into a little bit of perspective. So yes, go ahead. Absolutely worry about environmental impact of computing, but at least get it right.
Vugrincic: You're saying that AI-generated profile pictures are not the problem?
Allemang: No. You're going to ruin your environment a lot more by washing your shirt twice.
Vugrincic: So what is the future of AI?
Allemang: I'm very fond of an artist who goes by the pen name Claire Silber. She made the comment that (AI) is the future of human civilization. So on the one end, I would just love that for her to be right. You know, basically that the human race moves forward in intellect and society and everything, because we finally actually understand all of what we know. And then of course, on the sort of science-fictiony side, the computer starts becoming so autonomous that we're not sure why we're here anymore. And then you've got the utopian ones where the computers look after us and we don't have to bother with silly things like politics and government and feeding ourselves. I think all of those things are actually possible. And I don't know which of those directions we're going to find ourselves in.
Vugrincic: I think I'd be more worried if you said you knew for sure, so...
Allemang: You know what? Yeah, I would too!
