Slavi Kaloferov

Generative AI | Week 4 | Micro UX


Brief

Design a way for a person to pass as a ‘generative AI’ in an everyday setting

Teammates

3 teammates | Hong Zhou, Qibin Chen, Xiaole (Zoey) Zang

Timeframe

7 weeks | 20.04 – 08.06.2023

Development

Testing blender

Conclusion: Humans heavily rely on abstraction in their language and effectively communicate by leveraging the contextual understanding of the recipient.

Commands such as “Tell me how to create an aircraft” can be further deconstructed into smaller action chunks, which can then be broken down even further. However, computers still operate using abstractions of keys and images, which ultimately necessitates delving into the binary level of 0s and 1s.

We thought that everything can be translated over different mediums, exploring language, image, sound, and 3D as starting points. 

Concept for universal language

We faced some challenges related to literal translation, as encoding the information in a way that makes sense to humans often resulted in seemingly nonsensical or gibberish representations. Additionally, the process of encoding itself is subjective, as it was predetermined by us, which compromises the objectivity and universality of that language. 

While my coding skills have proven to be advantageous for the teams I’ve worked with, I have primarily taken on the role of a developer, leaving little room to expand my skills in research or design. In this particular setup, it would be inefficient to teach coding to someone else in order to free up my time to focus on other areas, given the significant gap in their coding experience and the level of proficiency required for our desired outcome. 

Teaching someone a skill I possess to handle tasks is appealing, as it would enable me to gain hands-on experience in the skills I am more interested in. Additionally, teaching skills are valuable in fostering a shared understanding between me and my “student,” enhancing collaboration and knowledge exchange.

In my opinion, the direction of utilizing 3D Clouds was the most promising, as it offered a high level of flexibility and the ability to express and visualize data types. However, we faced challenges in automating the process of translating images, sounds, or text into clouds, as it required the use of Blender software, which seemed very difficult or even impossible to automate.

Reflection from the Feedback

I realised that there are often underlying themes or larger concepts at play, which, when incorporated into a project, can result in something deep and meaningful. By attaching our project to these greater themes, we have the potential to create work that resonates on a profound level.