The Ambiguity


︎ Date: May, 2021
︎ Catergory: Generative Design, Machine Learning (GANs), Print
︎ Format: Interactive website & printed newspaper
︎ Tools in use: Runway ML, P5.js, Adobe Indesign, Adobe Illustrator, Adobe Photoshop


︎ Featured in LCC BA GMD 2021 Graduate Showcase


The Ambiguity Newspaper is an AI-generated newspaper that embraces the objectivity and subjectivity against the topic of the political status of Taiwan. The newspaper consists of 2 machine learning models to generate the ambiguous portraits and irrelevant text story. By integrating P5.js, a web-based medium, the project welcomes users to input their perspectives and receive an absurd and rational composition newspaper.

The political status of Taiwan is a complex and ambiguous topic in East Asia and beyond. The newspaper aims to deliver the context by combining user-input prompts and background knowledge. Users can generate an ambiguous portrait fitting the generated story that comes from their inserted perspective.

With 12 pages, each page has its own topic with provided background knowledge from Wikipedia. I asked 11 people with different nationalities and identities to interact with the assigned page and send it back to me. The final result of the printed copy is a collective perspective from both human and machine.




↑ The Ambiguity Newspaper Showcasing in Animation.


↑ The generated images from latent space trained by StyleGAN2 Machine Learning model.


↑ AI-generated in-between President Portraits. Training dataset is the mixture of Taiwan and China Presidents from Google Search.
 

↑ Screen recordings of users interacting on the newspaper website by inputting short prompts and generating the profile images.


↑ Printed Copy of The Ambiguity Newspaper. Hand model: Jacob Chung.


↑ Scatterd printed copy.