top of page

Micro UX - Week 2

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

Methods:

Team: Tanya Singh, Akriti Goel, Jakob Prufer, Ruoxi Song, Changlin Hou, myself

 

The first task of the week was to divide up into smaller groups. My personal interest was in the area of generative sound AI, for a number of reasons:

  • The most familiar form of AI is typed language, (eg. Chat GPT) which felt like we would be taking an easy route;

  • The medium of sound can be impactful and I was curious about how we (humans) detect and interpret different sounds.

  • Sound was a medium that I hadn’t worked with, and maybe selfishly, I was keen to use this project to expand my knowledge.

Luckily, some of our group were also keen, so we formed into a smaller group with the focus on this sound.


Further research was undertaken into what sound AI’s exist; the findings demonstrating a range of poor, clearly un-human voices to those more impressive; another set of AI-generated music, and a very small amount of any other sounds.


We also decided to look further into the brief for more clues or guidance; still trying to get our heads around it.


Dissecting the brief. Image by author.


It became evident that the brief wanted us to understand how people experienced AI as it currently exists in their lives; we needed to understand their perceptions and expectations before we could manipulate this to deceive them.

Research and planning. Image by author.


Reviewing the remaining 6 weeks, we decided to commit ourselves into splitting the project into two halves, essentially conducting a small research project prior to designing the final outcome. Our philosophy was that we couldn’t possibly trick people into believing that humans are AI, if we don’t know what people’s boundaries with AI already are.


It was a bold decision, and eventually unanimous, but one I argued for strongly as I believe in having a strong foundation to design from.


We decided to test out the limits to conversations with AI, using the method of body storming again to embody AI and find out what would happen with different scenarios that were all sound-based. A teammate Jakob took on the role of the AI, gradually moving from regular AI to a more human-like personality


AI sound bodystorming. Video by author.



We learned from these imitations and experiences that AI with human traits felt unnerving. This was useful to understand to frame our questions for the upcoming research.


Our presentation was comprised of the body storming video, and the project planning and research we intended to undertake, including an in-person survey, and hosting a workshop with sound-AI based activities. The tutors were happy with this approach, actively encouraging and supporting the research methods we intended to undertake.


While there were less design-based activities and outcomes this week, having a structure for the project, a direction to follow and positive feedback meant that we were able to move forwards confidently.



Comentarios


bottom of page