Saige
Research and design project | 7 week timescale | Role: Research strategy, researcher and designer
Brief:
Design a way for a person to pass as a generative AI in an everyday setting.
Generative AI is becoming more and more prevalent in our lives, and mostly, the concern is what happens when we can't tell the difference between human and AI, when we're expecting a human.
​
But what happens when its the other way around?
Research Strategy:
- Learn what the general public know and understand about AI and generative AI, through interviews and surveys
- Interact and play with different forms of generative AI to find its capabilities and weaknesses
-Pair the two understandings in different scenarios via workshops
Research Method 1: Bodystorming
Bodystorming is a fun, engaging way to get off the paper and into the physical space. Using props and creating mini-scenarios allows us to explore the boundaries of the concept, without getting too bogged-down in the finer details.
​
We used this to act out AI interactions in real time, testing how people felt and behaved with AI.
We learned that there is a line where AI starts to sound more human, but too human: when it becomes overfamiliar, suddenly, we're not keen on AI any more.
Research Method 2: Street Survey
The street survey was curated specifically with three 30-second questions for a stop-and-go with the general public. While Granary Square (London) was a small area, it had an eclectic mix of business people, shoppers, tourists and students, giving us a broad range of respondents.
​
The benefit of being in person for a survey is that gives opportunity to ask follow up questions, and to ask important questions about their answers - like why?
We did try offering a snack to encourage participation - surprisingly, it really wasn't necessary - AI was a topic of interest for many people. We found that the level of knowledge about AI was much lower than we expected, with many misconceptions or over-assumptions being made about its capabilities.
Research Summary
Through our interviews, surveys and testing, we had learned the traits that people associated with AI. ​
This became our guidance for the next stage of the design journey - putting it on its feet, and and testing the insights against designed scenarios.​
​
Research Method 3: Workshop
Having learned what AI can do, and differently what people think it can do, I hosted a workshop focussed on testing our insights drawn from research into generative AI and current perceptions of this tech.
Activity 1: Listening Activity
After hearing a lot within the primary research about devices like Siri and Alexa, we wanted to play with the idea of talking to AI. We played 14no. clips to the participants and asked them to note whether they sounded like AI, Human or they weren't sure. The purpose of this activity was to understand what voices and phrases would more likely be accepted as AI, and how believable AI voices might be. There was a fair amount on ambiguity or guessing, and a notable amount of AI being presumed to be human. The key findings were:​​​​
-
42% Average accurate guesses for either human or AI
-
26% Average instances where people couldn't make a decision
-
48% Instances where AI was identified as human
Knowing that there was only just over a 40% success rate in identifying human voices and phrases gave us a space
sit in the uncertainty for the design of the outcome. ​
​
Activity 2: Storyboard Exercise
Participants were asked to think about where AI exists in their lives and where they could envisage it becoming involved in our lives in the future. This helped situate the 'everyday settings' in AI might be more believable. Whilst there were some more extreme ideas, there was also believable places where AI could come in. This taught us that AI has the ability to seamlessly integrate within lives in ways we weren't expecting.
​
Activity 3: Scenario Exercise
In a group of three, participants were either assigned as a human, rehearsing to go on a date; or acting in collaboration to give AI-like answers. Participants acting as AI were given a 'cheat sheet' on AI characteristics to assist with their answers, which was derived from all the previous research.
​
Design and Play
Engaging in playful sketching and creating physical prototypes has proven to be a valuable approach in my design ideation process, especially when faced with abstract design briefs. Embracing a playful mindset encourages me to explore creative and unconventional avenues, providing a liberating space for my ideas to flow freely.
Sketching is a quick and low-pressure method for me to visually communicate concepts while crafting physical prototypes offers a tangible, hands-on experience that goes beyond traditional sketches.
This hands-on approach allows me to discover unforeseen possibilities and facilitates the development of innovative solutions. By infusing playfulness into my design process, I break away from conventional constraints, fostering an open-minded and experimental mindset.
Outcome
Have you ever been nervous about asking a roommate to move out? Or to ask someone out for a drink? Maybe, you need to have a conversation with your parents, about their health. ​What if you could practice this conversation, with a calibrated digital assistance, before the real thing? You'd get the chance to figure out the words you want to say, check how it might be perceived, before the real-life conversation. ​Meet sAIge, your home assistance for difficult conversations. Powered by generative AI and calibrated to your personal situation, you can test out things you're too afraid to say,
Reflection
Extra Bits
This brief was entirely baffling at the beginning. Not knowing enough about AI, and only knowing of its complexities was daunting, however the power of this project was in the research with people.
​
The pivotal moment was when we realised we weren't creating a whole new AI system, but instead, we were deceiving people, playing with their perceptions and leaning into their natural behaviours. The research gave us our grounding from which to design which ultimately made the project successful.
