Have you seen the viral video of humpback whales swimming in the Ala Wai Canal? Or what about snow on the top of Lēʻahi (Diamond Head)? You’d be among millions of viewers on social media.
Gabriel Yanagihara, a computer science teacher at ‘Iolani School, is behind the local viral videos made with generative artificial intelligence. The Conversation spoke with Yanagihara about the unconventional class assignment that led to the videos, and how we can all be more informed about the ever-evolving technology.
Interview Highlights
On what led to the videos, and how we can be more informed about generative AI
GABRIEL YANAGIHARA: Me, alongside a couple other teachers at ʻIolani School, we’re always looking at places where the students can kind of explore with new technologies, try new things in kind of like a, not like a walled garden, but like a safe place. So this summer, we did an AI boot camp… So one of the projects that spurred from that was the idea that an awareness of the necessity for AI literacy within our keiki and our kūpuna… One of the first examples from that that really kicked off was ideas like, we’re doing all this work to clean the Ala Wai Canal, what would it look like if it was actually so clean and a little bit deeper, right, and you could put a humpback whale in the Ala Wai Canal? So we went to the AI tools. We used Google’s Veo 2 new AI model, and we were able to put humpback whales in the Ala Wai Canal. And these photorealistic videos being posted and circulated online for millions of views of people who, even though it’s labeled AI, aren’t reading the caption and seeing that and thinking, ‘Oh my god, the Ala Wai is clean now. They’re humpback whales, right.’ That’s obviously not the case yet. But there’s two parts, right, one, it illustrates the need for being aware of what these tools can do. And two, that paints a pretty cool future that we can visualize, that could spur public action towards actually cleaning the Ala Wai.
On how people can determine if something is AI-generated
YANAGIHARA: So putting that image out there, right, there are a couple errors. Anytime you generate AI visuals, right, it’s not going to be super perfect yet, but things they noticed, right, a couple of checklists, like, did it come from a credible source? I’m just a random Instagram account, right, was it vetted? Do you see it by media channels? Is HPR publishing about it? Right, so checking your sources for it, and two, just having a critical sense of, is this practical? Is this realistic? Is this something that could actually happen even more so, diving deep, looking carefully at the video, and seeing that there’s two-way traffic on the Ala Wai Canal, and a bunch of people caught that, and that was, like, one of the key defining factors for a lot of people is like, ‘Oh, I caught that, that was AI, because I know there’s no two-way traffic on Ala Wai.’
On how people can utilize AI for good
YANAGIHARA: If I could get one video of me explaining this stuff to be as viral as someone like the humpback whale or putting snow on Diamond Head, like, if I could get it that popular, we could solve this problem overnight, we could get people aware of what it’s capable of, so that they don’t fall for the scams that might be coming in the coming years… I would say, like the main thing with this and any other emerging technology, just to stay curious, right. If you’re curious, you want to know answers. If you want to know answers, you’re going to dig into how the technology works… So as long as you’re curious, and you take the extra time it takes to make yourself educated and what they can do, you don’t have to be an expert in using them, but as long as you’re aware of what the capabilities are, you’re not going to fall for it, or you’re at least going to second check if someone sends you like a deepfake or something like that…