‘Cicer cum Caule’: Artificial Intelligence Meets Polish Pierogi
Interdisciplinary curator Anna Desponds and cultural technologist Philo van Kemenade discuss their project ‘Cicer cum Caule’ – a workshop aimed at explaining some fundamental notions of artificial intelligence through the process of making traditional Polish dumplings.
Philo van Kemenade (PvK): We are running a workshop called ‘Cicer cum Caule’. Shall we explain how it came about?
Anna Desponds (AD): Cicer cum Caule is an introduction to understanding artificial intelligence (AI), served in the form of an online pierogi-making workshop. The goal is to explore concepts related to AI, machine learning and datasets through the process of making the Polish dumplings. I first developed it as part of the #ArtHouse art residency at the Adam Mickiewicz Institute and then asked you to join as a technologist.
The workshop explores two essential skills for the (unknown) future, namely food making and technological literacy. Together with the participants, we make sense of technology through the process of preparing pierogi. The goal is to make the inner workings of AI more understandable, accessible and tangible, using Polish cuisine as an intuitive medium.
Picture display
standardowy [760 px]
Pierogi with fruit are the ultimate taste of Polish summer, photo: Anna Desponds
PvK: While many have been dreaming of (and some actively working on) artificial, human-level intelligence in humanoid shapes, little by little, we’ve grown into a reality where tiny little bits of AI are all around us.
In many ways, AI algorithms have already become ubiquitous: you may already rely on AI supporting you in making ordinary decisions throughout your daily life. These range from recommendations for movies on Netflix and books on Amazon, to friend suggestions on social media, as well as tailored advertisements from Google and auto-corrections in apps we use to communicate on a daily basis. The businessman and computer scientist Andrew Ng says that artificial intelligence is the next electricity.
The way algorithms work is often misunderstood and misrepresented in public debate. AI is an abstract, complex and varied research field, far from common sci-fi anthropomorphic representations of artificial intelligence. These metaphors lead to misunderstandings of what AI is and its capabilities. For example, have you ever read a headline about ‘An AI’ who has done all sorts of clever things? Referring to ‘an’ AI as if it was an entity not only falsely ascribes agency to algorithms, but also hides human agency and human labour
Because of these varied misunderstandings and misrepresentations of AI, it’s important to foster a wider understanding of what AI is, where it comes from, the ways in which we already encounter it in our daily lives, and the mechanisms behind it.
Picture display
standardowy [760 px]
Google search results for the term ‘AI’ suggest it is a blue humanoid, photo: Anna Desponds
Picture display
standardowy [760 px]
Slide from the ‘Cicer cum Caule’ workshop, photo: Anna Desponds
AD: During the Cicer cum Caule workshop, we try to make AI more understandable – using steps from my grandmother’s pierogi recipe as a metaphor.
PvK: But Anna, the name of our workshop sounds rather cryptic – does it help to make things more transparent?
AD: ‘Cicer cum caule’ is Latin for the Polish saying ‘groch z kapustą’, or hotchpotch – a mix of things that don’t belong together. The Latin translation of the saying was coined by the poet Julian Tuwim as a title for his essay series. The workshop bridges the multisensory physicality of food and the intangibility of technology. It mixes fruit with data, languages and narratives. And while the language of science is traditionally Latin, ours is food.
Picture display
standardowy [760 px]
‘Cicer cum Caule’ book by Julian Tuwim, photo: Anna Desponds
PvK: Looking at pierogi, AI and gender in technology all together definitely sounded like quite a mix! In Dutch, we call such a mishmash ‘ratjetoe’.
AD: Yes! The workshop also involves funny music, hand-drawn slides, and some cooking wizards. I am in a kitchen in Berlin, and you are on a boat in Amsterdam. The participants make food in their respective kitchens.
I think during the past months, a lot of spheres which were previously separate started to blend together. Kitchens have gained new significance. During the lockdowns, many of us have become domestic gods and goddesses as well as technology experts – the amount of Google searches regarding baking terms skyrocketed, and so did the stocks of Zoom Video Communications. Our professional and personal lives started merging. Someone told me that they have never seen as many strangers’ kitchens as on work-related video calls.
Picture display
standardowy [760 px]
The ‘Cicer cum Caule’ workshop comes with hand-drawn slides, photo: Anna Desponds
PvK: Did you become an expert on food and technology yourself while developing this experience?
AD: I definitely learned a lot... although when the pandemic hit, I already had a career as digital media curator, a Master’s in food studies and a part of a more than 150-year-old sourdough named Cornelius in my fridge. It helped me cope with this new personal and professional situation. But I am certainly not a technology expert.
I was born as a girl in Poland in the 1980s. I got my first computer when I was around 25 years old. I know a few languages, but all of them natural ones. I have been wondering: had I been born a boy, would I be making a project to understand AI through cooking? Or would I rather be trying to learn how to cook by using algorithms?
PvK: I think I can relate. I’m a guy, born during the 1980s in the Netherlands. When I started studying AI at the University of Amsterdam, just four out of 40 students were women. Starting my first full-time job at a Big Data startup in London, out of my 7 colleagues, only one was female, and she wasn’t part of the development team.
AD: According to a UN report published in 2019, women and girls are 25% less likely than men to know how to leverage digital technology for basic purposes, four times less likely to know how to programme computers and 13 times less likely to file for a technology patent.
But a male-dominated technology landscape doesn’t have to be a given, and there is interesting precedent to suggest the contrary. Ada Lovelace, a mathematician and writer who introduced many computer concepts, is considered the first computer programmer. She died in 1852, so more than 168 years ago.
PvK: Ada Lovelace was ahead of her time in many profound ways. She met the inventor and mathematician Charles Babbage when she was aged 17 and learned about his analytical engine. She was one of the few people who really saw its true potential. Lovelace didn’t just come up with entirely novel ideas about the ways in which calculating machines could be instructed to perform specific tasks. She was the first to foresee computation’s ability to emulate creative processes and engage in poetry and art.
Picture display
standardowy [760 px]
Portrait of mathematician and writer Ada Lovelace by William Henry Mote ,1838, photo: public domain
PvK: According to Wikipedia, the term ‘computer’ itself, which was in use from the early 17th century, meant ‘one who computes’. It designated a person performing mathematical calculations, not an object – before mechanical and later electronic computers became commercially available. From the second half of the 19th century, women became increasingly involved in this act of computing, and ‘computer’ came to carry a female connotation.
AD: And then women stopped coding. Among other writing on the subject, there is a very interesting episode of NPR’s podcast Planet Money that tries to explain this phenomenon. The amount of women studying computer science was rising alongside other disciplines dominated by males, such as medicine, law, and physics. At a certain point, while the involvement of women in the other fields continued to grow, the number of female computer science students dropped. Apparently, this is interconnected with the arrival of personal computers.
David Sims wrote, in a story for The Atlantic:
Text
As ubiquitous as they might be now, in the 1970s, few things were more mysterious and unknown than the “personal computer.” For years, these shadowy, ever-shrinking machines had been touted as the next revolution in the American home, although few people had a sense of how they might actually work. In April 1977, that changed with the launch of the Apple II, one of the first affordable, mass-produced PCs in history. Here was a machine small enough to fit in the home and intuitive enough to use without a programming degree. Still, the advertising challenge—how to convince people to shell out for a product no home had ever needed before—was daunting. The best answer was the simplest one: Make it seem like they’ve always been there.
Author
From 'How Apple and IBM Marketed the First Personal Computers' by David Sims, 'The Atlantic', 17 Jun 2015
These personal computers were advertised for men and boys. This ad for the Apple II computer says it all: while the man is exploring his new Apple II, the woman in the background is happily preparing food.
Picture display
standardowy [760 px]
A woman 'computer' by an IBM 403 Accounting Machine, born 1945, photo: Flickr Attribution 2.0 Generic (CC BY 2.0)
PvK: I had a computer as a kid, studied AI and today write code as part of my work. As much as I was happy to join this project from a technologist’s perspective, initially I strongly encouraged you to do it on your own. I believe running a workshop from an ‘outsider’ perspective is a powerful way to demystify AI and show how it can be approached.
AD: I took two online courses at the University of Helsinki. It’s a great programme created with the mindset of making AI accessible and understandable to everyone. I read and am still reading books, many of them related to bias in AI and technology. Some of the interesting writing include: Data Feminism by Catherine D’Ignazio and Lauren F. Klein, Algorithms of Oppression: How Search Engines Reinforce Racism by Safiya Umoja Noble, and Automating Inequality: How High-Tech Tools Profile, Police, and Punish the Poor by Virgina Eubanks. I’m not an expert, but I for sure know way more about AI than I did a year ago.
PvK: And thanks to you, I learned how to make pierogi! And got to know a few cheesy songs too.
AD: When I started the research, I was first considering focusing on the preservation of food – on ways to make it last until the end of the world. Cucumbers and cabbage pickled in salt are one of the most popular traditional Polish superfoods. They can last years.
When researching and prototyping Cicer cum Caule, I realised that not only was it generally challenging to explain the details of a recipe remotely, without being able to touch things, but also that many other things can go wrong in an online cooking class. During one of the prototyping sessions, someone’s cutting board caught on fire. This was an important lesson about safety. There could also be sanitary issues, and the jars of preserved cucumbers could have been rotting after the workshop, without me even knowing.
While my mum was my main culinary consultant, your suggestion was to look for a recipe with interchangeable ingredients. I ended up choosing vegan fruit pierogi. Fruit pierogi are very versatile, with ingredients easily accessible anywhere on the planet. I experimented with a range of fruity fillings. From all the fresh frozen and dried produce I’ve tried so far, only papaya hasn’t really worked for me.
We support the participants through the entire pierogi-preparation process, and at the end of the workshop, we all have a meal together.
PvK: And while making the pierogi, we talk about AI.
AD: Yes, the pierogi are a starting point for a conversation about machine learning. A friend of ours, Alexa Steinbrück, who works with creative AI, said:
Text
I like to explain algorithms as digital recipes for data manipulation, because that’s just what they are. You have ingredients like flower/fruit (which represent data) and then manipulations like hands/temperature/mixing (which represent functions), and then control structures like repetitions, etc. And as soon as you see algorithms this way, as sequences of instructions, they are defetishised.
Algorithms and recipes are a starting point for the conversation. Demystifying AI is an important goal of the project.
AI is frequently depicted in culture as a powerful human-like force, and often a woman – from movies like Her to voice assistants, such as Siri or Alexa. This technology is gendered as female and pays a submissive role. While technology has been and to a large extent still is a field dominated by men, pierogi have been traditionally made by women. We are using these historically female fields of activity to look at the patriarchal fields and narratives related to technologies of the future.
PvK: Food can be a great tool for visualisation. During the workshop, we talk about a variety of subjects related to machine learning, while using the ingredients in front of us to understand notions and processes — such as algorithms, supervised and unsupervised learning, and the frame problem in AI. We discuss the anthropomorphising of technology and how this hides human labour. Furthermore, we address biases in these technologies, which often reproduce and amplify inequalities which exist in the environment where they are created and utilised.
These issues have come under increasing scrutiny in recent years, not least from artists and designers engaging critically with the ways they arise and impact society.
Take, for example, the work by the artist and researcher Mimi Ọnụọha. In her work Us, Aggregated, she presents photos from her family’s personal collection alongside images that have been algorithmically categorised as ‘similar’ by Google’s algorithms. Who is in charge of the curation of this aggregated form of identity, which hangs together from both similarity and randomness?
Picture display
standardowy [760 px]
‘Us, Aggregated’ by Mimi Ọnụọha, photo: Anna Desponds
PvK: Another example is Excavating AI, in which Kate Crawford and Trevor Paglen examine how computer-vision research relies on data that is gendered, racialised, ableist and ageist. They dissect the biases found on ImageNet, a database of more than 14 million images, which are hand-annotated to indicate what objects are pictured. Note that the authors of the ImageNet project have recently recognised these issues and have started work to ‘illuminate the root causes of these concerns and take the first steps to mitigate them constructively’.
Picture display
standardowy [760 px]
PvK: AI has a variety of uses in everyday life, but also in the cultural field. In my work at the Netherlands Institute for Sound and Vision, I often make use of AI techniques to help users access vast amounts of collection items via intuitive representations such as speech transcripts or visual similarities. I’m excited about the virtuous circle that is starting to link these two worlds. AI is providing helpful techniques to make sense of large collections. It also supports the long-term conservation of cultural heritage.
Picture display
standardowy [760 px]
AD: Some of my favourite projects at the intersection of AI and culture include Noah Levenson’s Stealing Ur Feelings – an interactive, responsive story explaining how our emotions are being tracked by technology. Another example is In the Event of Moon Disaster by Francesca Panetta and Halsey Burgund, focusing on deep fakes and alternate history. The project consists of a fake video of President Richard Nixon delivering a historical yet never-used contingency speech, which was written in 1969 in case the moon landing was a disaster.
Picture display
standardowy [760 px]
PvK: This year in March, at the Mozilla Festival, I am co-curating the brand-new Creative AI space, where we’re collectively exploring how humans and computers can be more creative together to envision and question how artificial intelligence connects to our daily lives. This series of events will include collaborative art making, hands-on learning, open studio sessions, critical reflection, forward-facing discussions and web-native exhibitions.
PvK: Your idea to mix food and tech also reminded me of other projects — such as ‘Data Cuisine’, which explores data and information through the medium of food.
AD: Another reference was Kuchnia (Kitchen), a TV show in which Wiktor Niedzicki explains how physics work while sitting at a kitchen table. More than 100 episodes have been aired on Polish TVP since the 1980s. One of my all-time favourite documentaries is the Slovakian film Cooking History by Peter Kerekes. It tells the story of military conflicts through food. I general, I did quite a lot of research when preparing the project. I even looked into the food habits of tech billionaires.
For instance, the dumpling emoji was created as an act of empowerment and democratization: two friends discovered there was no dumpling emoji, and ended up changing how an emoji becomes an emoji. I hope this project helps — even if just a little bit — to further democratise technology and cooking, and to make them more accessible.
Conversation conducted in English, Oct 2020
Philo van Kemenade (The Netherlands) is a cultural technologist and educator from Amsterdam. He creates tools, stories and things in between to amplify human connection with the arts and culture. At the Netherlands Institute for Sound and Vision, he works on innovative user interfaces for audiovisual collections. Philo co-founded the digital arts & culture festival Sensorium and is an initiator of the Storytellers United network. He studied AI at the University of Amsterdam and Cognitive Computing at Goldsmiths College in London. In his free time, he enjoys cooking and eating Japanese pancakes.