Imagine a world where you love technology. You love it because it represents you. You were included in its creation. You have influence over its design and applications. You're not even the coder, but you know what the code can do.
If we want tech, and the products, services, and systems its intertwined with, to represent all of us, to work for all of us, it must be designed by all of us. Our goal is to provide radical access to emerging technologies like artificial intelligence and blockchain. We want people of all ages, races, professions, and genders to be able to use tech as a medium in their work.
How do we give people an embodied understanding of emerging technologies? Is it possible to create analog learning experiences that affect the design of technology? What tools will empower anyone to experiment with technologies the same way that they currently experiment with using different materials for mocking up products, different layouts for visual design, or different schematics for systems design?
Components of the Initiative
This initiative consists of two introductory learning experiences: (1) Designing with Machine Learning, and (2) Designing with Blockchain. Several tools are used within these learning experiences, and also stand alone. These include the I Love Algorithms Card Deck, Algorithm 'Mad Libs' and Data Sourcing worksheets, the This is Design Work and Map the Problem Space frameworks, and the My First Blockchain workbook.
Our Design Process
The last year spawned a number of new technologist-created tools that allow people to use and attempt to 'see inside' certain algorithms and models. But most are still highly technical, requiring prior coding knowledge. We approached the topic as designers that make things in many analog mediums. Our first experiments were with our community of 90-ish design instructors, and we've iterated through a range of audiences: K12 educators, middle and high school students, highly technical and non-technical Stanford students, educators and students from other universities around the world, and the general public. Along the way, we've shared our work publicly, and incorporated feedback from computer scientists and others.
People We're Reaching
Since launching our first prototype workshops in September, 2018, we've run more than 700 people, ages 11-80+ through learning experiences, both on campus and beyond. We know Stanford is a place of privilege, and made moves to remove barriers to attendance to on-campus events. In an effort to maximize access, we covered the cost of parking and local transport at our free public workshops if needed.
When we started this project, our hypothesis was that it would be most impactful for those that were tech novices. This happened, but we've been surprised by the reaction from technology experts. Having the tools to prototype the implications of their work seems to resonate. This quote from a recent participant, a CS graduate student, sums it up well: "I used to think machine learning was easy, now I think that applying it effectively is like rocket science."
I Love Algorithms Card Deck
This card deck explains six common machine learning algorithms: Classification, Clustering, Reinforcement Learning, Dimensionality Reduction, Regression, and Association. Each algorithm is explained in three ways: (1) Cartoon, (2) Simple text description (3) Styles of questions you might ask that the algorithm might service. We've done multiple iterations of this deck incorporating feedback from machine learning experts and a range of users. All workshop participants receive a deck and it is now available as a standalone resource too.
This is Design Work Poster
Design work is more than products, experiences, and systems. It also includes the technologies and data that power them above, and the implications of all of our work in the world in the near and long term. This poster shares our perspective of design work being an interconnected landscape.
Map the Problem Space Framework
This is a poster-size workshop tool that helps participants identify all the opportunities for design interventions across the data, tech, product, experience, system, implication landscape described above. Teams use it to expand the scope of their work.
Algorithm 'Mad Libs' and Data Sourcing Worksheets
I wonder if we could use ___(algorithm type)______ to ___(do or change what) _______? It'd be terrifying if we could use ___(algorithm type)______ to ___(do or change what) _______?
What gaps exist in your data? By using the data set you've identified, what does that mean you are prioritizing? Who or what are you excluding?
These worksheets are used as part of the Designing with Machine Learning Activity and guide teams through trying on different algorithms and sourcing data sets with an eye towards being aware of biases.
Designing with Machine Learning Workshop
This is a 90 minute learning experience where teams use human-centered design to work on real problems and learn to utilize machine learning to scale their concepts. The workshop is approachable by people with any type of technical or design background. It starts with having the whole room act out a classification algorithm using the 'hot dogs or legs' meme, and uses it as a way to highlight what an algorithm is and how it makes decisions. It's also a way to show everyone that each of us shows up in our work, and that we all have varying perspectives and biases.
The main workshop activity is a design challenge that varies (ex. re-design the college admission process for first-gen students or re-design the city parking experience for ride-share drivers). Teams first map the problem space of their work using the 'Map the Problem Space' poster. Then, they identify a user and create a boutique concept that would be special for that individual. Next, they use the I Love Algorithms Card Deck, Algorithm 'Mad Libs' and Data Sourcing worksheets to scale their ideas for 10,000 users. Finally, they use some futures tools to imagine both the potential positive and negative implications of their work two or more years into the future.
Designing with Blockchain Workshop
We kick off this workshop with an activity where the whole group acts out both a centralized and a distributed system. Using Polaroid cameras, markers and paper participants create a blockchain in the physical space. Within minutes, participants understand blockchain to be a tool that can disrupt the many systems in our lives (and not just cryptocurrency).
The bulk of the 90 minute workshop is the My First Blockchain Activity. This exercise has groups deconstruct a system, identify intermediaries, eliminate them, explore trust and incentives, and eventually identify what would be recorded on a blockchain that might replace a key intermediary in their system. Like the machine learning workshop, it also ends with a focus on future implications. This activity is released under a Creative Commons License (Attribution Non-Commercial). Some of this exercise was inspired by conversation and an activity first run by Engin Erdogan, Burak Arian, and Cenk Dolek. The d.school version has been since adapted and run by Laura McBain and Kwaku Aning, as well as Jennifer Gaspar-Santos, and Alexander Jones for varying K-12 audiences.