HYPERCINEMA - NYU ITP


Duban Morales
Date 10/23/2024

Synthetic Media Discovery
Daniel Ambrosi "Milford Sound, Fiordland National Park, New Zealand"




Questions - Why do you consider this synthetic media? What can you find about how this was made (if you can)? If it’s machine learning based can you find information about the data set used to train it? What are the ethical ramifications of this specific example (if any)? "Milford Sound, Fiordland National Park, New Zealand" uses artificial intelligence (AI) to modify or enhance a photograph in ways that go beyond traditional editing techniques. Ambrosi frequently works with deep neural networks—AI models like DeepDream—to introduce intricate, surreal patterns into natural landscapes. The result is a blend of human creativity and machine algorithms, creating a visual style that feels like a mix of both the organic and the artificial.

In this image, you can still recognize the iconic landscape of Milford Sound, but it’s layered with fractal-like structures and complex, surreal textures. Ambrosi's approach is what’s called computational photography. He starts by capturing large-format photos of landscapes, and then runs them through neural networks like DeepDream, which are specifically designed to amplify certain patterns and create surreal effects.

From what I’ve gathered, Ambrosi likely uses a customized version of DeepDream, an AI developed by Google to detect and intensify patterns within an image. The neural network essentially “hallucinates” details—like patterns, faces, or shapes—that give the final image its dreamlike, fractal quality.

Though the exact datasets Ambrosi uses aren’t always clear, DeepDream models typically rely on huge datasets like ImageNet. This includes millions of categorized images that help the AI learn how to recognize and interpret visual patterns when processing new images.

Works like this raise interesting questions. By using AI to manipulate landscapes, the line between what's real and what's been digitally altered starts to blur. For landscape photography, this might make people question how much of nature is being represented accurately. If someone looks at this without knowing it’s been modified with AI, they might get a distorted sense of what the actual location looks like.

In Ambrosi’s case, it feels like he’s using AI as a collaborator—guiding the process but still letting the machine play a significant role. The ethical concern, though, is where to draw the line when the AI starts making more creative choices than the artist.