AIxDESIGN
[2022] YEAR IN REVIEW /
STIMULERINGSFONDS REPORT
TABLE OF CONTENTS
THANK YOU
[CORE TEAM]
[CONTRIBUTORS]
[SPEAKERS]�Whales for Climate
Aarati Akkapeddi�Emily Martinez�Andreas Refsgaard�Maya Man
Felipe Sanchez-Luna�Soyun Park
[EDITORS/CO-AUTHORS]
As we reflect on the past year, we are filled with immense pride and gratitude for everything we have accomplished together. None of our achievements would have been possible without the trust, and support of our community members, especially those named below.
This community is the backbone of AIxD, and we are truly humbled by the impact we have been able to create as a result of our collaboration.
And a special thank you to Stimuleringsfonds! Your financial contribution have provided the necessary resources to implement essential programs while adhering to fair pay standards. Your belief in our mission and willingness to invest in our organization have played a pivotal role in our success.
MEETING STIMULERINGSFONDS OBJECTIVES
This chapter addresses how the 2022 program of AIxDESIGN addresses the conditions, criteria, and objectives of the grant as stated in the Digital Culture Grant Scheme document.
1/ The project was initiated by Nadia Piet, digital research, creative technologist, and founder of AIxDESIGN.
2/ This project contributed to the crossover of digital culture and other design disciplines. Its audience consisted of people and organisations are or are interested in using AI in a cultural or design-led context:
A] THIS PROJECT PROMOTED THE DEVELOPMENT OF ARTISTIC QUALITY
In PROGRAM 1: AI PLAYGROUND, we showcased 4 accomplished artists using AI in their creative practice to help demonstrate what’s possible across four creative disciplines – image-based art, text-based art, sound-based art, and body-based art.
Since program end, we’ve been commissioned by SubLab – the R&D team of production studio Submarine Channel – to continue this work, showcasing how artists use AI to in their motion-graphics practice. Learn more about this work at < STORY&CODE >.
B] THIS PROJECT STIMULATED EXPERIMENT AND RESEARCH
In all programs, we experimented with how we might apply open-source principles (most often found in software engineering) to a design research project. Whenever possible, we publicly documented all project approaches, and research findings. These tools and resources are published under a Creative Commons license including all workshops materials and recordings so that other might reuse them in their own design discipline. These resources have used in design and arts schools in the NL, UK, Germany.
Most recently, John Meada, VP of Design and Artificial Intelligence at Microsoft, recommended AIxD’s AI Ideation Cards in his annual report, Design in Tech 2023: Design and AI.
In PROGRAM 3, MAPPING AI AND DESIGN, we explored how participatory research could increase the diversity of perspectives into the study of AI and design disciplines. In what is typically a conservation dominate by the Global North, this research benefited from the research and insights of 70 contributors from 28 countries from around the world.
This research approach has informed a new design research project commissioned by Sound&Vision, led by Nadia Piet, to explore how AI might be used to in motion-picture and sounds archives. Learn more about this work at here.
In PROGRAM 4: AIXDESIGN LAB, we experimented with AI tools in the workflow of professional graphics and motion designers. We published the results of 3 experiments with our research logs.
Since program end, we’ve been commissioned by SubLab – the R&D team of production studio Submarine Channel – to experiment with how AI might be used in film making. Learn more about this work at < STORY&CODE >.
Project amai! used AIxD’s AI Ideation Cards in co-creation sessions with government, companies, academia, 'middenveld' and citizens to ideate new AI-solutions for a societal impact. By the end of 2023, 3 projects will be granted € 125.000 to continue development.
AIxD has been a very interesting space to discuss my art practice with other artists and creative technologists. Can’t wait to see how how this community grows and evolves!
– ERYK PETERS, MEDIA ARTIST
C] THIS PROJECT SUPPORTED REFLECTION AND DEBATE
AIxDESIGN is a remote-first, Amsterdam-based community of designers, researchers, creative technologists, and activists using AI in their pursuit of creativity, justice, and joy.
We are a living-lab making art and sharing alternative AI narratives to those in Silicon Valley, peddling corporate AI. We also make and hold space for our global community of 8,000 to host and engage in critical and hopeful conversations about the design and use of AI in creative industries both in-real-life (IRL) and online across quarterly community events, our Slack, Linkedin, and Instagram Live.
Specifically in PROGRAM 2: EVERYDAY DATA (H)ACTIVISM, we welcomed 17 non-technical professionals from across 5 continents to reflect on the impacts of AI bias on our collective creative potential. This program was recruited via an Open Call in which we received 50 applicants from 20 countries (150% more than program capacity). The results of this work was publicly published to encourage further reflection and debate.
Since program end, we have been invited to speak about this work at events hosted by cultural and educational institutions like Dutch Design Week, Impact Hub, University of Arts London, Central Saint Martin (UK); and Reshape Forum for AI in Art and Design (Germany).
Everyday Data (H)Activism sparked critical thinking, and active investigation across our cohort. From this reflection and debate, we created a library of resources, a wealth of information that can be leveraged to educate others about responsible data and AI use.
– VERONICA CHEN, DESIGNER
EXECUTIVE SUMMARY
2022 was a year of unbridled experimentation. From our projects to our ways of working, we tried to imagine what it might look like to be proud of our work – both our journey and output. We’re immensely proud of what we’ve achieved with our time, budget, and team.
For more information about our budget and how it was spent, see the accompanying document: 2022 Financial Report for Stimuleringsfonds.
In PROGRAM 1: AI PLAYGROUND, we explored how we might use publicly available AI-driven tools to fuel our design and art practices.
In PROGRAM 2: EVERYDAY DATA (H)ACTIVISM we explored how we, people without “technical” skills, can practically and critically engage with and protest against algorithmic systems.
In PROGRAM 3: MAPPING AI AND DESIGN we defined this emerging interdisciplinary field in an effort to co-create and document a common language so we can reflect and debate our experiences and insights.
In PROGRAM 4: AIxDESIGN LAB we explored, as designers, how we might use AI-driven tools to simplify our workflows and serve as a creative partner.
AI Playground applies the philosophy of Anji Play to AI. In this program, hosting artist talks, hands-on workshops, and collating guides from our learnings, we explore AI tools in an unstructured manner with a focus on friendship over competition, messiness over perfection, process over completion, reflections over portfolio building. This sentiment is beautifully captured in this video created by project co-lead @computational_mama.
Curated by Nadia Piet & @computational_mama, this season of AI Playground included four modules that explored the use of AI to create image-, text-, body-, and sound-based art and were attended by 400+ creatives.
Each module consisted of three parts. Each module began with an ARTIST TALK. An invited guest shared a behind-the-screens look at their work and how AI/ML plays a role in their creative process.
In a follow-up WORKSHOP, @computational_mama and a expert guest hosted a beginner’s workshop on the artist’s subject area, sharing helpful tips, tricks, and tools to get started.
Each module wrapped with a written guide, AIxD’S GUIDE TO ____, to share insights, reflections and resources - written together with program collaborators, and published by the AIxD team.
OUTPUTS
<01> AI & [MOVING] IMAGE
Humans have always made images, driven by a need to preserve memories and tell stories. Yet, with each image-making innovation – like photography and video – we are forced us to rethink how we make and use images. The emergence of AI and ML tools invites us again to re-imagine our role as both image-makers and image-consumers.
1.1 ARTIST TALK // Whales for Climate Change
1.2 WORKSHOP // Good Old D(AI)s with Aarati Akkapeddi
1.3 AIXD’S GUIDE TO AI x [Moving] Image by Catarina Rodrigues
<02> AI & TEXT
Chat-GPT has become a household name. In this module we explore what is legitimately hyped and not for text-generating AI and how designers and artists are using this tool in their work.
2.1 ARTIST TALK // Recl(AI)ming Pleasure with @Queer AI
2.2 WORKSHOP // N(AI)ve Poetry Collaborations with Andreas Refsgaard
2.3 AIXD’S GUIDE TO AI x Text by Fred Wordie
<03> AI & THE HUMAN BODY
Until recently, human pose detection and recognition technology was not accessible to non-techies. In this module we explore how artists currently use this AI/ML application in their creative pursuits.
3.1 ARTIST TALK // Navigating Self & Body on the Internet with Maya Man
3.2 WORKSHOP // Learn to: Fingerp(ai)nt with words with @computational_mama
3.3 AIXD’S GUIDE TO AI x the Human Body by Cailean Finn
<04> AI & SOUND
In this module, we explore AI's application in music creation, the distinction between symbolic-based and audio-based AI, available tools for AI-audio experimentation, and the importance of transcultural AI.
4.1 ARTIST TALK // Building Soundscapes Through (Machine) Learned Histories with Felipe Sanchez-Luna
4.2 WORKSHOP // Composing Chrom(AI)tic Resonances: Turn Images Into Short Musical Scores with Soyun Park
4.3 AIXD’S GUIDE TO AI-generated Sound by Yaboi Hanoi
EVERYDAY DATA (H)ACTIVISM is a community-led research project that (re)imagines methods of everyday resistance using data in ways that are both playful and power-conscious. Led by Abdelrahman Hassan, this program consists of a series of three collaborative workshops designed to help us better understand:
SESSION 1
What are the realms of activism as it relates to big data and AI?
SESSION 2
What does data activism look like today? How do these tactics measure up against and alongside each other?
SESSION 3
How can we best prototype and maintain an open-source guide to Everyday Data Activism?
OUTPUTS
Everyday Data (H)Activism Toolkit
Everyday Data (H)Activism is a community research project that explores how ordinary people can engage in everyday acts of data activism. This toolkit is the output of this project, and acts as a living archive of data resistance.
See the Toolkit here >>>
Workshop Assets
We have published under the Creative Commons license the workshop plan, Miro board, and outputs for each of the workshop. We encourage the use and reuse of these workshops as a whole or in parts.
See Workshop Assets here >>>
Everyday Data (H)Activism felt like belonging to a communal garden: analyzing differents types of data-driven systems, identifying how AI can threaten human rights and then getting our hands dirty by creating concrete actions to address specific issues of algorithmic discrimination
As a designer, this project challenged my practice, with ideas from a rich mix of participants I would have never met if not for AIxD.
– JOHANNA MUSCH, CRITICAL MAKER
Led by Nadia Piet, this research project aims to map the emerging field of people, practices, education, and professional opportunities at the intersection of AI & Design.
Think of it as cartography project: an effort to map the intersection and entanglements between AI and design. It should offer an accessible and inviting answer to the question:
How do AI and design intersect (from a design POV)?
We explore this in two ways:
1/ AI for Design: How does AI impact design?
2/ Design for AI: How might design(ers) shape AI?
This provides a foundational mapping of the not-yet-defined AIxDesign field and creates a share language / understanding of the different practices within it.
It should help anyone with an interest to navigate the space, where they may want to position themselves within it, and have starting points to continue on their learning journey. This could be helpful for younger professionals, but also organizations to discover collaborators, or identify new opportunities.
For more information about this project’s multi-methodology approach, see here >>>
OUTPUTS
AI and Design Ecosystem
Within this project, we used Miro as a tool to gather our own research, collect community input/feedback, and to present the findings in a report-ish style. This “report” has been intentionally designed to look like a living document instead of a “final report”, because it is! We hope this in-flux / non-linear / mind-map presentation allows people to explore the information in ways that make sense to them, and encourages readers to add their own insights and experiences.
See Report here >>>
Listen to our collaborators discuss their findings here >>>
Interviews on Instagram Live
As part of our research, we interviewed inspiring practitioners working in AI and Design. In these conversations we discussed the motivations behind their work and mapped the ecosystems and intersections of AI & Design as they saw it. We discussed specific educational programs, career opportunities, and research topics they were personally excited about in the AI and Design space.
See our collection of Instagram Lives here >>>
The AIxDesign Lab is a space where we experiment with generative design & AI tools to support our own creative process and content creation process. The results are implemented into the AIxD communication channels, event setups, and internal team flows. We live-document our approach and learnings in the links below – the good, the bad, and the ugly!
Experimenting with AI Tools
Led by Yasmin Morgan, this project took a broad survey of publicly-available AI tools - like GPT-3 and charRNNs (text models), image generation models - to better understand how can practically use them in our day-to-day work. Philosophically, these experiments investigate the flattening of identities, bodies and aesthetics by large models (ie. GPT-3, DALL-E, etc) and ways of obfuscating / rebuilding these tools.
Building a Generative Design System
PART 1: Building a Design System
Led by Andrew Heirons, we developed a traditional design system that served as the crucial foundation for the generative tools we prototyped later in the year. Through this sequential approach, we ensured our generative design tools aligned with our organization's brand identity.
PART 2: Creating Generative Design Tools for that Design System
Led by Deniz Kurt with Design Systems International (DSI) we built our first generative design tool – Generative Flyer Maker (for Instagram). We’re excited to continue to create generative design tools for our most requested graphics, merging creative coding, generative design, and AI, to explore the boundaries of creative automation. So much more to come!
OUTPUTS
Experimenting with AI Tools
1 // AI Icebreaker Toolkit >>> by Yasmin Morgan
We use these AI Icebreakers during workshops & interactive sessions to instigate playful critically, demonstrate algorithmic bias, and let people experience rather than to merely explain certain computational concepts. This is an ongoing collection - some curated from the web, and some developed in-house at AIxDESIGN.
2 // Fluent Hallucinations, the Journey of Making a Tagline Generator >>> by Yasmin Morgan
Yasmin set out to examine the qualities of various text-generation models by creating a tagline generator to write up a resonating brand statement for AIxDESIGN. They considered the shift from older models like RNNs to transformer models as a transition from craft to synthesis while seeking a balance between believability and imagination in the generated text. The blog highlights the potential of AI-generated text and sparked discussions on the nature of human-like imitation versus the unique creativity of machine-generated content.
3 // Using Image Segmentation to Make Generative Images of Playgrounds >>> by Yasmin Morgan
Yasmin ran an experiment around the creation of a custom COCO dataset of annotated playground images for the purpose of generating interesting algorithmically-generated backgrounds for the gather space we use during the AI Playground program using image segmentation, GANs, and the SPADE COCO dataset. The goal was to explore the annotation of images, the creative applications of image segmentation, and the ethical considerations related to dataset preparation.
4 // Tool for Pixel Collage-y Project Graphics for AIxD >>> by Marina Cardoso
Working with Marina Cardoso, a creative coder and graphic designer, we created a custom generate tool to create project graphics. Inspired by a tool from Constraint Systems and their open-source code shared in GitHub, It creates collages based on a pixel complexity sorting algorithm that layers a handful of input images into a visual assemblage.
Building a Generative Design System
PART 1: Building a Design System
We built a design system that serves as our foundation to help us easily and quickly experiment with AI design tools in the future.
PART 2: Creating Generative Design Tools for that Design System
With our design system in place, we are ready to create a suite of generative design tools like this one >>> made DSI’s Mechanic tool that helps us design AIxD event flyers in a flash!
APPENDIX
PROPOSED v. REALISED
Changes between proposed and delivered activities have been detailed in the table below.
1/ AI PLAYGROUND | |||
| PROPOSED | CHANGE | REASON |
1.1 | Each module will offer 3 parts: Showcase; Playground; how-to-guide | No change | |
1.2 | Module topics will include: Body as data ; Audio-related, Hearing, Music; Text-related; Image-related | No change | |
1.3 | At program end, we will organize one in-person event together with Creative Coding Utrecht | No in-person event | As this was the first time running this program, we had under-budgeted program activities. Also our co-host @computational_mama and many of our artist speakers were located internationally so we cut this. |
1.4 | The talks and workshops will be hosted hybrid or fully online. | All talks, workshops were hosted online. | Same reasons as above. |
1.5 | The talks will be €4 and the workshop €6. We’ll offer a program “membership” for €25 euros for access to all the sessions. | No fee charge for talks and workshops. | This fee was estimated for in-person events not online ones. We wanted to make it as accessible as possible so did it on donation-basis. We did not host in-person workshops so we did not charge a fee. |
1.6 | The sessions will be recorded and (partially) published on YouTube and IGTV. Written and simplified tutorials will be published through Medium. | No change | |
2/ EVERYDAY DATA (H)ACTIVISM | |||
| PROPOSED | CHANGE | REASON |
1.1 | 3x online workshops hosted between March - May 2022 | Workshops were held in Sept - Oct 2022 | Workshops Preferred dates for Lead and participants |
1.2 | We’ll announce an open call to recruit 15 data practitioners, data researchers, and community organisers to serve as researchers | No change | |
1.3 | Working sessions will be online, with option to host last session offline. | No change | Last session was hosted online to maintain equal access for our international contributors. |
1.4 | Sessions will be documented in Miro | No change | |
1.5 | Outcome of the three sessions will be a prototype of a card deck that is available digitally and in print. | No change | Cards will be printed after testing has ended, estimated Dec 2023. |
1.6 | Researchers will be reimbursed for their contributions | No change | |
1.7 | We will ask participants to write about their experience contributing to the project. | | |
APPENDIX
PROPOSED v. REALISED
Changes between proposed and delivered activities have been detailed in the table below.
3/ MAPPING AI AND DESIGN | |||
| PROPOSED | CHANGE | REASON |
1.1 | The research will be conducted throughout the year, and a report will be issued at the conclusion of the year in November/December. | Report was delivered May 2023 | Due to changes in the team including project leads (changed 2 times), we’ve had a delay on the project. This has been communicated with Stimuleringsfonds which allowed us to extend the deadline to end of Feb. |
The final report was delivered as an interactive Miro board, not a static PDF. | 1. To create efficiencies in project delivery, we minimised gap between doing and delivering the research. 2. As an emerging field, it’s more valuable to present our findings in an representatively “messy” way, as a living document, instead of a (over-)simplified, “final” version. | ||
1.2 | The research will include methods such as desk research, content analysis, qualitative interviews, and generative workshops. | We also conducted participatory research with 70 AI and/or design practitioners | We wanted to experiment with participatory research approaches to test how they impact research findings. This experiment was a success; we will continue to deploy these methods in our creative research practice |
1.3 | We will conduct interviews with ±8 practitioners & experts to inform research. The first and final episodes of the podcast will be with the AIxDesign and research team, sharing their research intentions and final insights. These interviews will be edited and published as a series of short-form podcasts with episodes dropping bi-weekly from April until October. | We hosted 4 “podcasts” on Instagram Live | Instead of recording and producing a podcast for distribution Spotify, we hosted these conversations on Instagram Live: 1. Instagram Live offers better accessibility for hosts and guests because we didn’t have to buy or install specialised technology and programs. This also meant it was simpler to build flexibility into our podcasts so that different people could plan, host, and publish. 2. Instagram Live offered an existing publishing infrastructure that our existing community could easily engage with. |
4/ AIxDESIGN LAB | |||
| PROPOSED | CHANGE | REASON |
1.1 | We will experiment with generative design and AI tools for our graphic design needs | No change | |
1.2 | We will develop a (generative) design system using Mechanic, a design system tool developed by Design Systems International. | We first had to build a design system. This helped us create generative tools aligned to that existing design system. | We faced a critical blocker – AIxD didn’t have their design system documented in order to code a generative design system. That required us to re-focus our efforts on creating the foundational blocks required for a generative design system. |
1.3 | We will explore how we might use AI/ML tools in other AIxD activities, e.g. interview summaries or newsletter creation. | No change | |
1.4 | We will share our learnings from these experiments - explaining how to use them, and honestly speaking on their abilities and limitations. | No change | |