Digital Clone


︎ Date: May, 2021
︎ Catergory: Speculative Design, Generative Design, Interaction Design, 3D Modelling
︎ Format: Audio-activated Installation
︎ Tools in use: P5.JS, Blender, Adobe After Effect, Adobe Photoshop


︎ Featured in LCC BA GMD 2021 Graduate Showcase
︎ Featured in UAL 2021 Graduate Showcase



Digital Clone is an audio-activated installation as an intervention for the speculative scenario in the year 2036. We will be living a life full of the Internet of Things (IoT), Digital Twin infrastructures, and 5G connection, which will become intimately integrated into our daily life via the internet.

This project aims to provoke the awareness of digital privacy through our daily devices, which have been always listening to our day-to-day conversation to pick up useful keywords for ads dataset, for example, Google and Facebook.

We are, at the same time, being updated to the ‘Digital Clone’ which tries to portray and approximate who we are in the form of digits, as if we are cloning different selves when using, interacting, or talking to personal devices.


For the installaion proposal ➔ Click HERE to the PDF deck.




↑ An audio-activated interactive installation that provokes a digital privacy discussion about the constantly listening features from our daily devices and smart objects. Check the video on Vimeo if the player is not working.

The p5.js prototype below recognises the keywords from your conversation, which the keywords are extraxted from my instagram data.

Try to say a sentence composed with keywords listed in the ‘Ad Interests List.’ You have to allow the browser to access your microphone first. The black area will disply the keyword one by one as you speak. Give it a try & Have fun.


↓ The following p5 prototype can be opened here if it is not responding. Some browsers may not support the embedment feature.

↑ Especially thanks to LCC’s Creative Techinology Lab Technician Edmund Oetgen.


↑ How the installation works, from recording the surrounding to displaying the animation according to the keyword.


↑ Using t-SNE to locate the keywords in 3D space and how the location distorts the 3D model.


↑ Different distortions based on the keyword's location in the 3D space.

↑ How P5.js recognises the keyword from normal conversation.