Through new co-creation and speculative design methodologies, the LocAI team explores the role of local context in AI product design.
By Daniel Goddemeyer and Rob Marchant
Daniel Goddemeyer is an independent designer and consultant who works with companies to envision future products that utilize the potential of data- and AI-driven technologies. He also teaches workshops using experiential methods that inspire product development with more nuanced, future-facing human perspectives. Rob Marchant is a program manager at Google AI, working on future-facing AI design projects.
The Challenge: AI and Cultural Context
AI-powered products are increasingly being deployed at global scale to support us in our everyday lives — from suggesting songs to helping us complete our text messages. Over time, many of these systems improve in their ability to support specific needs of individual users.
All users are different however, and once deployed, these products must adapt to unique individuals in a variety of situations and use cases that are specific to the user’s local environment — with its unique norms, values, belief systems and etiquettes.
In order to enable AIs that are highly mindful of cultural specifics, we need new methods to create more direct connections between users, AI designers and developers, and we need to raise awareness of local cultures throughout the product development process.
The LocAI Project
LocAI is a co-creation workshop series that investigates this challenge — particularly around the role of cultural context when designing future AI technologies.
It is aimed at design students and guides them through a process from research to concepting to design and, through the resulting work, highlights how local attitudes, aspirations and needs are being projected onto technology.
Participants are tasked with identifying possible use cases for AI assistance that exist within their immediate surroundings. They must then visualize these local AI futures (aiming for a timeframe of the next five years) and transform them into active learnings for the development of future assistive AIs.
With these exercises, LocAI aims to create an open, shareable methodology that — as a set of tools — allows designers to explore the use cases of an assistive AI within immediate local contexts and, through this, create a more direct connection between cultural specifics and global development.
Case Study: LocAI Singapore
We piloted LocAI in September 2019, with a group of students from Nanyang Polytechnic in Singapore.
Through a series of quick, successive sprints over the course of two weeks, the LocAI team, including Matt Jones, Principal Designer at Google AI, guided them through a design-focused AI research and design process. By focusing on exploring future possibilities for AI assistants that could support specific, local needs, students created future-facing design solutions that demonstrate how AI could be embedded into the everyday lives of Singaporeans.
The LocAI Process
01 Understanding the context
To get an initial feel for what distinguishes everyday life in Singapore, students started by identifying the distinctive local demographics, professions, use cases, quirks and issues.
Next, they spent a day researching in the field to observe and explore the group, topic or issue they were specifically interested in.
02 Understanding the technology
After their initial field research, the students were guided through a series of exercises to understand the present workings and future potentials of AI.
The data we create
To grasp how our own data is the raw material for machine learning algorithms, we asked the students to take out their phones and draw the most used apps on their home screens.
For each app, they were asked to list the interactions they perform with it and the corresponding data that they think they leave behind through using the app.
Students then had to call one of their parents live in class to see how they would define ‘AI’. This helped to get a feel for the diverse associations we have with AI, often heavily influenced by popular culture — and come to a unified definition to work with.
It provided a fun starting point, and helped to differentiate between fact and fiction — between realistic capabilities in the coming years and science fiction.
“The part where we all of a sudden had to call our friends or family to ask what is AI was my favourite activity. It was a pretty wild but interesting experience!” — Workshop participant
Knowing me, knowing you
Once the class arrived at a shared definition of what we commonly understand as AI, we followed up with an expert talk that equipped students with an initial understanding of the present and future potentials of AI.
Students then used this knowledge to build on the data portraits their fellow students had drafted.
Imagining they had a fellow student’s data for one year, they speculated what insights around their behaviors, interactions and daily routines could be derived from such data.
They then presented what they might learn about that student back to the class, “as seen” through the lens of their personal data and AI.
03 Applying the technology through design concepts
Following a discussion around the nature of personal assistants and assistive intelligence in their local surroundings, students then started concepting what first use cases of an intelligent personal assistant in their immediate surroundings could be.
“Coming up with something for your younger self was an interesting exercise to get the brain juice flowing. Fun that there were no restrictions.” — Workshop participant
Following a fun thought starter exercise where the students had to develop an AI for their parents to assist them in dealing with the students at their most annoying age, they then considered how a future, intelligent AI would be able to support their researched user group, and translate this into conceptual directions for future features and products.
In order to quickly test their hypothesis and manifest their ideas visually, students went through a quick exercise of creating one-page adverts, visualizing their concept and defining its main features.
04 Prototyping concepts
Prototyping their ideas without any ‘actual’ AI at their disposal, students had to be creative in how they could let the class experience their ideas.
Simulating Intelligence through role playing, building quick mock-ups and involving their fellow students in live experiments, the initial ways of testing their ideas gave them quick insights and initial feedback.
05 Final projects
Finalizing their projects, students then developed visual scenarios, product mockups and images that exemplified how a future AI assistant would practically play out in the local context of Singapore.
From an AI supporting Singapore’s unique culture of tolerance between religions, ethnicities and demographics, to creating support systems for the interactions at local Hawker (Food & Drink Court) centers, to an AI supporting Singapore’s large population of migrant workers — their visual stories sparked new thoughts and discussions around an AI-assisted future in Singapore.
Although the students focused on their immediate surroundings, some of the resulting ideas such as an AI incentivizing recycling, or an intelligent stroller assisting elderly populations, were broadly applicable beyond the immediate context of Singapore.
Reflection & Learnings
Running this initial workshop in Singapore, as with any pilot, some things worked better than others, and we certainly left with ideas for improvement, from extending the user research phase to adding new exercises.
Feedback from the students was extremely positive and the students left with new perspectives around the capabilities of AI and how these future AIs could assist people in their immediate surroundings.
Looking beyond this pilot, the LocAI project aims to continue to develop a replicable, participatory methodology that can allow anyone to explore the role of future AI assistants in their direct local context. Not only in a classroom-led scenario like in the pilot, but directly to designers and developers wherever they are.
Our hope is that LocAI creates new connections between local context and global product development and contributes to AI products that are more mindful of local aspirations, issues, hopes, needs and use cases.