Annual Conference 2021 of the Collaborative Research Centre 1187 “Media of Cooperation”, University of Siegen 25–29 October 2021

Carolin Gerlitz, Deputy Speaker and Principal Investigator of the CRC Media of Cooperation, opened the conference by situating practices of learning in a digitalised world, pre and during Covid-19, framing „Study with me“-videos as a platformed way to share co-presence in learning situations as an embodied experience despite physical distance. On the flipside of this Gerlitz describes the non-human side of learning based on data, processed by algorithms that learn user’s practices and adjust content and practices of learning on a global scale. With these different dimensions of learning in mind she introduced the conference topics and handed over to Jutta Wiesemann and Martin Zillinger for the introduction of Jean Lave, keynote speaker for the conference opening.

30 Years of Situated Learning

The CRC has always been inspired by non-media centered media research and sees all media appear as cooperatively developed cooperation conditions, the practices and techniques of which arise from the mutual production and provision of common means and processes.  Thus, Lave’s work is highly relevant to the CRC because of her ethnographic focus on relationships of learning that highlights pathways to the investigation of new communities of practice that revolve around new media technology or algorithms.

Lave’s landmark work „Learning and Everyday Life“ from 1991 turns 30 this year. The keynote given on the 2021 annual conference of the CRC invites her and us to think about the things that have changed and things that have not changed since the initial publication. Is the key argument of “situated learning” (Lave 2019, 133-157) still adequate to investigate a new datafied society, new technologies and their practices?

Three things that have not changed since the publication of “Situated Learning”:

1. Ethnographic research of craft apprenticeship is still Jean Lave’s first grounding as an anthropologist doing ethnographic work.
2. The Ethnographic method was not ever only ethnographic, but also always ethnographic/theoretical, hence, constantly theorising ethnographic work.
3. “Situated learning“ was a personal and critical-political project at the same time, which might be the reason for the popularity of the book, according to Lave.

Denmark’s Production Schools

Lave’s recent research was focussed on learning in Danish Production Schools. Target group (and community of practice) “Production Schools” were young people who despised and were alienated from school – and from capitalism. Their educational practice constituted a contradictory but productive relationship between production and education – labour and learning.

2017, the system changed: production schools were, as Lave puts it, „swallowed up“ in a new system, emphasising classical classroom teaching. Craft masters were fired and replaced with teachers with less practical experience. The entire system’s space, time and personal structures were transformed to emulate formal education. Two different understandings of what it means ‚to learn‘ were waged against each other: a hegemonic and a counterhegemonic perspective.

Image Credit: Ministry of Children and Education (2020): Production schools, Copenhagen, https://eng.uvm.dk/upper-secondary-education/production-schools

Production Schools:

– workshops where young people work and learn at the same time
– apprenticeship for 6 months to 1.5 years, arrival and departure at no specific times
– vision for young people: Gain a stronger grasp on their own future possibilities through collaborate engagements in their work

For more information visit: https://www.schooleducationgateway.eu/en/pub/resources/toolkitsforschools/detail.cfm?n=7026

„Hegemony as a relation between the present and the future“ (Stefan Kipfer)

To conceptualize hegemony and thereby hegemonic / counterhegemonic perspectives on learning, Lave highlighted the quote by Stefan Kipfer (2008). Through his work, Kipfer connected Gramsci’s and Lefebvre’s works on “dialectical relational method of inquiry”, which also inspired Jean Lave’s thinking about hegemonic and counterhegemonic understandings of learning: In the Production-Shool-Workshops, the young people found access to mundane practices (from cooking to metal work) and the possibility to participate in a cooperative structure. These conditions were produced by the production schools with the workshop masters as teachers and “not-teachers” at the same time; learning was seen as a transformational change. Lave stresses that production schools „do not separate learning to do something and doing something” from each other and consequently: „you can only learn something if you‘re engaged in doing it“. The production schools therefore gave young people the outlook to change their relations of alienation in a process of learning for everyday life. Yet the goal always was to produce something of value (not strictly monetaristic).

For the participants this meant that the process of learning was also an opportunity to produce themselves as producers of increasing value, while gaining more comprehensive skills. This enabled them to gain a greater perspective and a larger grasp of ways in which that genre of production is part of larger political and economic systems, granting them, as Lave puts it, unquantifiable transformational experiences. This unquantifiability however stands in contradiction to the formal educational system making the schools susceptible to reformation. As Lave remarks later: „What is hegemony? It is capitalism engaged in trying to reproduce itself “, making the reorientation of the production school system an expression of renewed alienation to people who did not seem to fit into the formal education system in the first place.

Against the backdrop of this inspiring case for situated learning practices and its tragic demise, Lave concludes with a more broad outlook on practices of learning: Historical and political dimensions shape how learning is enacted and framed and therefore should play a role in research inquiries. Especially this final remark was picked up in the debate following Lave’s Keynote as participants talked about the conditions for academic learning and teaching and especially the lack of situatedness and engagement in actually doing research and enjoying it. Researchers nowadays often only work from paper to paper, deadline to deadline, losing the possibilities for deeper thinking, interactive learning and doing research.

References:

Kipfer, Stefan (2008): „How Lefebvre urbanized Gramsci: hegenomy, everyday life, and difference“, In: Goonewardena, Kaniska/Kipfer, Stefan/Milgrom, Richard/Schmid, Christian (eds.): Space, Difference, Everyday Life. Reading Henri Lefebvre. New York: Routledge.

Lave, Jean (2019): Learning and Everyday Life. Access, Participation, and Changing Practice. Cambridge: Cambridge University Press.

>click here to enlarge image<

Graphic recording of the opening keynote by Jean Lave, Oct 25th, 2021.

Graphic recording by: Coline Robin
https://coline.graphics

“Anticipating Parting” – a brief audiovision in anticipation of Pip Hare’s presentation “Practising Being an Indian Soldier” on Oct 26, as part of panel 1 “Intercorporeality and Learning”

“What is ‘learned by body’ is not something that one has, (…) but something that one is.” (Bourdieu 1990: 73). Pip Hare will present camera ethnographic audio-visual material that makes embodied learning visible: a 3 ½ year-old child and his mother interact with various media to co-operatively perform ‘being an Indian soldier’ with mimetic practices including packing and preparing, taking leave, travelling, advancing, firing and fighting, climbing, and flag-hoisting. Meanings are co-constructed and identities incorporated as the participants practise an imagined future.

Rachel Chen received a BCNM 2021 Summer Research Award. Hear about her great work developing a platform for non-speaking children on the Autism spectrum.

This summer I collected data for my dissertation, a Design-Based-Research (DBR) doctoral study that examines an embodied-design technological platform developed for non-speaking children on the Autism spectrum. The study combines my background in Linguistics, examining the naturally-occurring embodied interactions of Autistic individuals, with a mixed-methodology approach through my Ph.D in Special Education. The platform, the Magical Musical Mat, maps touch to music, allowing users to explore music together through high-5s, claps, and different touch-based gestures.

In the summer of 2021, I was in Singapore for fieldwork. I brought the Magical Musical Mat, into the homes of Autistic children. I worked on the prototype in between sessions with my participants, iterating on the design of the mat in between. At each session, I took video recordings of my participants interacting on the mat, and also interviewed the children’s parents. I then customized the mat and the sounds that we placed in the mat to the preferences of the children.

The grant helped in paying for prototyping materials, such as yoga mats, arduinos, fabrics, sensors, and also helped me cover transportation costs required to travel from one participant’s home to another. Conducting in-person research in the midst of Singapore’s changing COVID-9 regulations meant that unforeseen costs such as disinfecting protocols needed to be considered. The summer grant helped us cover these unforeseen, miscellaneous costs.

In addition to completing data collection for my dissertation, I presented my work at the Interaction Design and Children conference, where I discussed preliminary results and future plans for my work.

First published: Berkley Center for New Media (2021) under the title “Summer Research: Rachel Chen on Embodied Platforms for Children on the Autism Spectrum” on Sept 1^st, 2021

This video is an introduction to prototype 3 of the Magical Musical Mat (MMM). The MMM is being co-designed by Rachel Chen and Arianna Ninh at University of California, Berkeley. Credit: Rachel Chen (2019): Magical Musical Mat, YouTube video, 11.11.2019, https://www.youtube.com/watch?v=VSWmGKWE0Vk

Magical Musical Mat
Being in Touch With the Core of Social Interaction: Embodied-Design for the Nonverbal, Ignite Grant

For more information please see:
https://magicalmusicalmat.com/

SLEEPING – SHOOTING – STIMMING

Starting a panel with a presentation entitled “Learning to (fall a-)sleep” sounds rather risky, but as it turned out, Larissa Schindler’s (auto)-ethnographic study of how babies learn to fall asleep was the opposite of sleep-inducing.  She pointed out that the apparently mundane and often taken-for-granted human practice of falling asleep cannot be wilfully “done” but is a highly culturally accomplished practice that must be learned.

Because it cannot be done, techniques are developed to create conditions that increase the likelihood of falling asleep, involving media that Larissa Schindler calls the “infrastructure of sleeping”: pacifiers, slings, night lights, music, etc., are deployed to help the baby fall asleep at a place and time deemed appropriate. They have to become accustomed to the sleeping habits of their families and societies.

Larissa Schindler invites us to follow her into the deeper layers of the seemingly obvious, as she shows how what seems “natural” at first sight is actually the result of an elaborate mutual learning process in everyday family life in which parents also learn to sleep anew.

In her contribution “Practising being an Indian soldier”, Pip Hare contrasted the supposedly peaceful scenario of learning to fall asleep with images that some audience members found disturbing, such as a mother and her son shooting at each other with toy guns. She showed two camera ethnographic videos of a 3 ½ year old boy, Abir, who, together with his mother, enacts practices of being a soldier in the Indian army. The whole living room is temporarily transformed into a virtual military training camp. Even the ethnographer, Pip Hare, is handed a gun so that she can participate in the ongoing performance.

A strange, if not uncomfortable, tension arises between aspects such as the mother’s and child’s affectionate gestures, the cute rucksack he fills with army figures, and the small red bike that Abir rides to go “in the army” on the one hand, and, on the other, the meanings brought forth in the situation: practising, and performing for the ethnographer, being an Indian soldier.

Pip Hare related her observational material to Bourdieu’s concept of habitus, asking whether the presented scenes can be seen as mimetic identification.

After showing the second sequence, in which Abir watches and acts along with scenes from a patriotic Indian war movie, Pip Hare proposed seeing the role of the movie less as a model than as a co-participant, and suggested that “embodied spectatorship” (Laura Marks) might be formative in engendering what Pip Hare then called “embodied citizenship”. It strikes me that this could be taken further: perhaps the sense of ambivalence or unease that some of the panel audience, as well as Pip Hare herself, described feeling as they watched the observational material, can be seen as reflective of their own embodied citizenship/personal biographies, which shape how they experience watching Abir’s and his mother’s actions.

Rachel Chen’s presentation “The sound of touch” also focused on the body as a medium of embodied learning. In her research with persons on the Autism spectrum, Rachel Chen found that non-speaking Autistic people’s bodily interactions are far more intense and intentional than most people realise. Electronic devices that are designed to help Autistic people communicate are often developed on the basis of a symbolic linguistic approach that does not correspond to the ways Autistic people actually communicate. Hence, they produce asymmetrical communication situations for persons who already find themselves in asymmetrical power relations in much of their lives.

Rachel Chen chooses a different approach: Drawing on a phenomenological concept of intercorporeality (Merleau-Ponty) and taking the repetitive movements of Autistic people, such as stimming, rocking, and flapping, as “key dimensions of Autistic socialisation” (Sinclair) – rather than as non-normative behaviours to be discouraged – Rachel Chen worked with a colleague to develop a magical musical mat that makes sounds when two people, standing on the mats, touch each other’s bodies. The mat uses the affective power of touch and sound to encourage Autistic and non-Autistic people to communicate non-verbally.

As different as the three presentations were, they all focused on the “body as active center of cognition, social understanding, and cultural making” (C. Meyer), on the socio-materiality of learning, and on what Jean Lave calls “situated learning”. Larissa Schindler, Pip Hare, and Rachel Chen vividly showed that learning is not a mental operation, but a social and cooperative practice in which human and non-human actors are mutually involved.

References:

Marks, Laura U. (2020). The skin of the film. Duke University Press.

Merleau-Ponty, Maurice (1945/1955). Phänomenologie der Wahrnehmung. Übers. von Rudolf Boehm. Berlin 1966 (frz. 1945).

Sinclair, Jim (2012). Loud hands – autistic people speaking. Washington, DC, Autistic Press.

Encounters. Now and then

Thomas Alkemeyer’s chapter “Befähigen” [Enabling] (in Pierre Bourdieu: Pädagogische Lektüren, 2016) was a guiding inspiration for me when I embarked upon my ethnographic research within the Early Childhood and Smartphone project in the CRC ‘Media of Cooperation’.  It was enabling for me at the time, helping me to recognize how participants enable each other to ‘play along’ as they acknowledge each other’s competences and join in as co-players. Keeping this in mind orientated me as I engaged reflexively with my diverse camera ethnographic observations.

So, as you can imagine, I was very much looking forward to Thomas Alkemeyer’s keynote speech at the conference.

With his concept of a ‘conditional agency’ that takes shape in the encounter of human and non-human participants (including artifacts such as tools, and technologies, etc.) in social practices, Thomas Alkemeyer articulates a critique of both the classical idea of an autonomous subject as the sovereign center of initiative and a (post-)structuralist understanding of subjectivity as a mere effect of pre-existing social structures or practices.

So how does his conceptualization of learning as an integral part of all kinds of practices, as a diffuse event, as enablement to play along ‘competently’ relate to his understanding of subjectivation? That was what I hoped he would elaborate upon in his keynote speech.

Praxeologically, Thomas Alkemeyer stated, learning can be understood as the successive incorporation of a repertoire of disparate dispositions that include movements, bodily techniques, and skills, as well as attitudes, inclinations, dispositions, preferences, and desires. It is only in encounters that one can learn how to do things to appear as a ‘competent co-player’.

Musing on this insight, I found my thoughts drifting… How might these ideas be applied to the current ongoing online encounter? Who is enabled to present themselves as a competent co-player here? But I couldn’t allow myself to reflect further, for fear of missing the next twist of the tale (and failing to act as a competent blogger). So, I turned my attention back to the lecture.

Encounters, I learned, are not only potentially enabling, they can also disenable. Learning may be fostered or hindered by the conditions that shape it. Emphasizing that the ‘capacity to play along is a relational capacity’, Thomas Alkemeyer defined processes of resonance and dissonance between incorporated dispositions and the objective circumstances as conditions that ‘enable or disenable’.

Indeed, I thought, the conditions that characterize attending a lecture via Zoom in the evening after several hours of online conference are not particularly well disposed to increase my capacity for attention. But then the narrative drew me back in with the next slide:
‘The emergence of something new in the encounter.’ That sounded promising, and Thomas Alkemeyer delivered what he promised, offering a very tangible example from the history of sport to illustrate how specific conditions enabled the development of a novel technique which, while resonating with traditional techniques, was also markedly different.

The Forsbury Flop – now the most widely used high-jumping technique – went on to revolutionize the world of high-jumping after Dick Forsbury used it to almost beat the world record at the Olympics in Mexico in 1968.

A number of diverse factors – technological-material, biographical, biophysical, and more – came together to produce the conditions that enabled the ‘invention’ and success of this new jumping style. Thicker, softer foam mats had recently been introduced, which enabled an athlete to land on their back without risking injury; Forsbury was a mechanical physics student and able to use his technical knowledge to analyze and improve jumping techniques; he was only moderately successful as a high jumper, which led him to seek alternative ways to meet his ambitions despite his athletic limitations; he found a coach who took him seriously and helped him to develop his technique. Last but not least, the Zeitgeist of 1968 in the USA was open to novelty and revolution in all aspects of society. Mass media celebrated Forsbury as a rebel who had broken through the stubborn conservative traditions of professional sport.

Thomas Alkemeyer argued that taking a praxeological perspective when considering dimensions of the social that are classically understood in relation to the concept of the subject – e.g., agency, mental perspectives, self-determination, critical self-correction – enables us to recognize that people, objects, and artifacts are more than just interaction partners or supporting elements present during particular actions; they are co-participants that enable each other to ‚play along‘ in their encounters, but can also disenable each other.

“Learning is not limited to intellectual exchanges between teachers and learners, rather, it is brought forth through the embodied entanglement of humans with the material-symbolic world.” (Alkemeyer and Buschmann 2016, quotation translated by Pip Hare). That was the sentence that I borrowed and drew upon in my chapter “Willst du mitgucken? Intervenieren – arrangieren – etwas sichtbar machen” [Do you want to look too? Intervening – arranging – making something visible] to help me analyze my observations of a situation in which a 1 ½ year old child, her mother, and a smartphone all co-participated to enable the child to practice playing a competent co-player.

>click here to enlarge image<

Graphic recording of Larissa Schindler’s talk on „Learning to (Fall A-)sleep“, as part of panel I „Intercorporeality and Learning“ on Oct 26, 2021.

>click here to enlarge image<

Graphic recording of Pip Hare’s talk on „Practising Being an Indian Soldier“, as part of panel I „Intercorporeality and Learning“ on Oct 26, 2021.

>click here to enlarge image<

Graphic recording of Rachel Chen’s talk on „The Sound of Touch: Designing for Intercorporeal Attunement with Non-Speaking Autistic Children“, as part of panel I „Intercorporeality and Learning“ on Oct 26, 2021.

Graphic recording by Coline Robin
https://coline.graphics/

>click here to enlarge image<

Graphic recording of Thomas Alkemeyer’s keynote on „Encounters. Praxeological Reflections on the Constitution of Contingent Agency“, as part of panel I „Intercorporeality and Learning“ on Oct 26, 2021.

Graphic recording by Coline Robin
https://coline.graphics/

How do we approach the field in a disarming and appropriate way?

After reading the panel’s title, I am wondering if there even is such a thing as „decolonising learning“. I think, we can all agree that we can decolonise teaching and approaches, but learning is an inherent and intrinsic action. We can decolonise narratives that form societal structures and patterns that would potentially shape how we perceive and learn. But can we decolonise learning, as bluntly as it is put here?

Two weeks ago, I heard someone at an event on disruptions in ethnographic fieldwork saying „Our heads are colonised.“ I think this is true and we have to face that. Not necessarily embrace, as this is not something we should cherish or value, but admit, face, and challenge. And yes, that is indeed uncomfortable. If it’s not uncomfortable, you are probably not doing it right.

As part of the Collaborative Research Center 1187 – ‚Media of Cooperation‘ annual conference about ‚re-situating learning‘, this panel gave space to provocative presentations of reflective ethnographic and playful approaches, appropriate collaboration tools and archives, and the roles of museums in making (hi-)stories more collaborative.

Nina ter Laan and Marike Minnema invited us to take part in their reenactment of the issues they face while conducting remote theatre activities in the computer club in the High Atlas Mountains. I was one of the researchers who set up this infrastructure, back in 2017, together with dear colleagues of mine. Witnessing how it grows, nurtured by the local community, is heartwarming. After some provocative thoughts from the audience, questioning this approach and comparing it to what the French did as the colonial power over North Africa, Marike concludes:

„But also, especially in a post colonial power, dynamic context, I always allow the participants to ask questions, to first interview the researcher. So who are you? You always get the questions. Are you married? You have children? Why are you here? What makes you come to Africa. And then we start doing theatre work that’s more poetic. And because it’s more poetic, it allows to see whatever emerges.“

Anne Weibert presents a participatory card game, which was developed within a computer club in Germany, mainly visited and appropriated by women with a migration background. The game explores everyday life with and without technology, carried out by fictional characters in ordinary settings.

„Our game is still young, but we have already learned a couple of things. We have learned how pictures are important. Our game gets by with very few written words. This creates a broad basis for understanding even cross language barriers, playing cards are sometimes blank, our game is expandable, this puts valued design sovereignty in the hands of the participants.“

Both these presentations, as well as the comment given by Badiha Nahhass touch upon questions I have been asking myself lately. What are methods and modalities that are more natural, authentic and appropriate for the people I work with?

How are we planning to deal with museums and the idea of „decolonisation“? Is collaboration key?

Janine Prins addresses the gap between cultural institutions and museums. To be more precise, reaching out to younger audiences seems to be a challenge for some museums. Coming in as a researcher, Janine observed, trying to shed light on what sometimes happens in the black box of co-creation activities.

Personally, I am pretty undecided of my position towards museums. However, turning them into places of genuine co-creation and allowing people to connect with what they have lost or even never really have, is an inspiring perspective.

„And I also remembered a visit to a museum in Marrakech, Morocco, with young Dutch-Moroccan students who wanted to get in touch literally, with their cultural heritage, the lives their parents led, which had gone lost due to migration. And they really wanted to hands on touch the material culture and not see it behind glass in the museum.“

In her presentation titled „The collaborative museum: Why it needs epistemologies; why it needs theories“, Michi Knecht proposes a number of points how museums could become more collaborative spaces, without collaboration becoming a ‘legitimization machine.’ To help museums deal with their colonial pasts and making different futures, she proposed 3 pre-requisites, including the recognition of the violence and destructive force of colonial powers and neocolonial knowledge systems that still exist today, the acknowledgment of multiple knowledges, and the joint definition of all those involved of what need to be done.

„Why does the collaborative ethnographic museum needs theory in order to change and to ask relevant question about the future of museums in the global context?“

Anna Brus, discussant of this slot, picks up on the notion of collaboration as an interactive and reciprocal practice. How do you turn collaborative activities within a museum into another exhibition that again can be explored, appropriated, and modified by museum goers?

How about we turn museums into spaces that provoke us, challenge us, make us question stories instead of giving us the impression of „knowing“? We already consume all the time. We should use our energy and turn it into action.

The closing words of this slot, and I just want to quote Janine Prins, are: „We have to stop being the white saviour.“

How do we archive and visualise collaboration?

Mike Fortun and Tim Schütz present an open source platform for experimental and collaborative ethnography. Mike outlines his presentation introducing collaboration in six different ways: „as an object, collaboration in theory, the politics of collaboration, what it means to design for collaboration, and then to practice it. And then finally, collaborations to come as an open ended, open invitation.“ This platform template has been used and adapted to a variety of ethnographic endeavours, allowing researchers and activists to collect, share, exhibit snippets of their work.

“And finally, that the ethnographic knowledge is shaped by the form and the genre in which it is carried.” I do very much agree with this statement, however I am wondering if this applies to both the process and the result. And, more importantly: Where does that leave us when working with communities whose communication modalities are oral? How do we do them justice?

How do we transfer ethnographically gained knowledge to people who did not witness it personally? This question is driving the research, presented by Volker Wulf and Konstantin Aal. Both are researchers in the area of Human Computer Interaction and Socioinformatics.They introduce the underlying framework of Grounded Design which calls for ethnographic and qualitative approaches to ground emerging technology within existing social practices and structures. „We believe that the quality of design of IT artefacts can be only understood if we understand how the social practices are changing due to the appropriation of these artefacts.“

Bridging the overall concept back to our research in Morocco that Nina und Marike talked about, they open the question of how to collect and archive stories and knowledge in a, for the local community, accessible and appropriate way. We have not found a solution yet, but from what I have learned so far, is that we constantly need to put an effort into questioning our ideas of how such a archive could look like in order to be not colonial about this.

The discussant for this slot, Matthias Harbeck brings us all a bit back from dreaming and hoping, and more into the reality of ethnographic fieldwork, that is often unplanned and spontaneous and demands flexibility and enduring tensions – also applying to technologies we wish or wish not to use.

„You have to use things like WhatsApp, which is a commercial product where you have judicial issues, ethical issues and using those softwares and also technical questions on how to get the transferred items out of those commercial platforms into i your open source platforms and from there further on. So I think that is something we researchers need to be aware of.“

The role of personal archives. A demand? A suggestion? A question?

Koen Leurs finishes the day with a keynote, titled „The Techno-Politics of Mobility and Immobility. An Archival Approach“, reflecting upon the perspectives and interactions with digital technologies by ‚people on the move‘. While I could go on about this talk, which was amazingly well put together, I want to emphasise one aspect.

„Over the last couple of years, I’ve started to reflect on the role of personal digital archives such as smartphones, personal profile pages on social media, as a means of eliciting reflections eliciting narratives of meaning making.“

I am asking myself: Am I treating my devices and pages as archives? A colleague and I often make fun of my (screen) desktop chaos and me being „a digital messy“. Well, I never felt the need to cherish my phone’s ability to store photos and written notes as I have not been forced to escape a war.

Koen introduces us to Patricia, a teenage girl from Aleppo, who fled the war in Syria. Patricia manipulated a photo of her baby sister, adding herself into it. „It was one of the pictures dear to her heart. For her it signifies both the period of living and separation from her loved ones.“ Her archive does not only allow storing, but manipulating. Manipulation here does not necessarily represent change, but desire and connected emotions.

Personally, I value Koen’s perspective on archives as spaces to make one’s own connections and create one’s own representations. However

„When we study technologies and the media, it’s important to look at how power is always there, ambiguous. So there’s power and but there’s also contestation inherently built into any technology. Infrastructure and the imaginary promise of connected presence is very much a gendered and radicalised unbalance to geography, political economy. And when we think about connectivity, let us think about how this connectivity is always almost instant, what happens in between and this can be at any moment.“

We cannot decolonise learning. We can decolonise our heads, our understanding of what is and what should be for people whose lives we will never live ourselves. We can and should, however, seek ways of allowing processes of genuine appropriation to happen. This is where we, as researchers, anthropologists, and designers of media and IT come back into the game. And it is our responsibility to do the job right.

>click here to enlarge image<

Graphic recording of the talk by Nina ter Laan & Marike Minnema on „Theater as Ethnography in Morocco. Making Meaning of the Digital Age Through (Online) Storytelling“, as part of panel II „Decolonizing Learning, Rethinking Research: Datafication and the Quest for Co-operation in Ethnographic Fieldwork“ on Oct 27, 2021. Graphic recording by Coline Robin https://coline.graphics/

>click here to enlarge image<

Graphic recording of Anne Weibert’s talk on „Utopia with (Out) Technology“, as part of panel II „Decolonizing Learning, Rethinking Research: Datafication and the Quest for Co-operation in Ethnographic Fieldwork“ on Oct 27, 2021. Graphic recording by Coline Robin https://coline.graphics/

>click here to enlarge image<

Graphic recording of the talk by Janine Prins on „Where the Digital and the Material Meet Up“, as part of panel II „Decolonizing Learning, Rethinking Research: Datafication and the Quest for Co-operation in Ethnographic Fieldwork“ on Oct 27, 2021. Graphic recording by Coline Robin https://coline.graphics/

>click here to enlarge image<

Graphic recording of the talk by Michi Knecht on „The collaborative museum: Why it needs epistemologies; why it needs theories“, as part of panel II „Decolonizing Learning, Rethinking Research: Datafication and the Quest for Co-operation in Ethnographic Fieldwork“ on Oct 27, 2021. Graphic recording by Coline Robin https://coline.graphics/

>click here to enlarge image<

Graphic recording of the talk by Mike Fortun, Kim Fortun & Tim Schütz on „Experiments in Ethnographic Collaboration. Purposes, Designs, Infrastructures, Invitations“, as part of panel II „Decolonizing Learning, Rethinking Research: Datafication and the Quest for Co-operation in Ethnographic Fieldwork“ on Oct 27, 2021. Graphic recording by Coline Robin https://coline.graphics/

>click here to enlarge image<

Graphic recording of the talk by Konstantin Aal & Volker Wulf on „Sharing Contextualized Knowledge: Curating Design Case Studies in the context of an E-Portfolio“, as part of panel II „Decolonizing Learning, Rethinking Research: Datafication and the Quest for Co-operation in Ethnographic Fieldwork“ on Oct 27, 2021. Graphic recording by Coline Robin https://coline.graphics/

>click here to enlarge image<

Graphic recording of the keynote by Koen Leurs on „The Techno-Politics of Mobility and Immobility. An Archival Approach“, as part of panel II „Decolonizing Learning, Rethinking Research: Datafication and the Quest for Co-operation in Ethnographic Fieldwork“ on Oct 27, 2021.

Graphic recording by Coline Robin
https://coline.graphics/

Cross-community learning is an important element in technology research and design. Too often, the element of learning collaboratively between researchers, designers, and target group representatives is treated rather implicitly. This often leads to important opportunities for innovation not being exploited.

In the following blog post, Katka Cerna, a collaborator in the EU project ACCESS, describes how learning spaces between technology researchers and older people as application partners are consciously used and reflected upon. She also describes particular challenges and requirements for cross-community learning with older adults. (Claudia Müller)

Katka Cerna, 26. August 2021

Aging Research during a Pandemic

We live in an aging society – with increasing longevity and improvements of care, globally, people live longer. However, a longer life does not automatically mean a more satisfied life. Many older adults suffer from an increasing isolation, lack of caring (both professional and personal) relationships. Digital technologies hold the potential to overcome some of these challenges, however currently also introduce a range of problems for this particular population. That is mainly because the new digital tools also require many older adults to learn a completely new set of skills connected to digital literacy. It is not hence strange that CSCW has increasingly become interested in this growing problem – after all, it is work to make the digital tools work for oneself (and not the other way round).

In this small series of posts, I introduce the different challenges that we faced when we chose to go online with older adults. More specifically I tackle the different aspects of working with the older adults when using Zoom within a participatory design project, namely the issue of digital ecologies, the issue of scaffolding, and the issue of self-directed learning; as well as will scratch out some of the solutions, both more specific such as didactic prototypes or broader visions such as meta-design space for older adults. Some of results are already published in the following text: Making online participatory design work: Understanding the digital ecologies of older adults.

Aging research in CSCW often takes place in the form of hands-on activities, where people come together in a room and together can make things, write on papers, show to each other what they do with different digital devices. However, the aging research during the past two years has been heavily impacted by the COVID-19 global pandemics, where older adults were often viewed as one of the most vulnerable populations. In combination with the spreading of COVID-19 is connected to in-person contact, it became crucial for us as CSCW and HCI researchers to explore methods which help us to keep on working and co-creating together with the older adults without actually meeting them in person.

We have addressed this challenge in our project ACCESS. Project ACCESS is an interdisciplinary and multinational project which focuses on fostering older adults’ digital literacy. The project is coordinated by Ass. Prof. Claudia Müller (group IT for the Aging society) from the Information systems and New Media department (University of Siegen). The more specific subproject which is presented in this blog post focuses more specifically on how to support the learning of older adults to use digital technologies through their involvement in participatory design process. A result of the project is a mobile demo-kit, which aims to provide learning resources to older adults and other relevant stakeholders about the experience and learning to use digital technologies. Current state of the demokit can be found here.

Practically, we have planned to organize a series of workshops, during which we wanted to together with the older participants explore a range of off-the-shelves technologies and how to use them. Our twenty participants come both from a local senior computer club and previous projects. They are a very heterogeneous group: they differ in age (65 to 88), digital literacy but also motivation to join our project. During our first session in person, we established a Telegram group (a messenger application), which we used to coordinate with the group.

However, as the pandemic progressed, we had to make a choice – either to stop with our activities or rethink them and move them online. In line with our action-oriented approach, we have chosen the later. We have tested several tools with the groups of our participants and ended up using Zoom. The choice of the tools (and any tools used later on) have been steered by the main goals of the project – to enable older adults to become more autonomous when it comes to using the digital tools in their daily lives. Hence, a lot of hands-on and collaborative exploration as an approach has been used as a way to foster self-directed learning of our older participants. However, that was not a simple task – none of our older participants knew how to use Zoom and we also had no previous experience on how to appropriately support them.

The first sessions were intense – the way Zoom is designed did not match the way our older participants were used to interacting with each other and we all had to learn a new way how to talk and listen to each other just so that communication with each other is possible.

About the ACCESS project. ACCESS project is an interdisciplinary and multinational EU project, funded by the JPI “More years better lives”. The project aims at fostering digital literacy of older adults by combining the perspective of lifelong learning of older people with technologies that meaningfully support everyday life. More information here.

Katka Cerna, PhD, is a researcher interested in learning and participatory design currently within the sustainability area. She uses ethnographic and participatory approaches to understand learning and design to enable people to take the actions they need in the more-than-human world.

Credit: Cerna, Katka (2021): „Aging Research during a Pandemic“, Online: https://www.eusset.eu/2021/08/26/aging-research-during-a-pandemic/

Future workshop between HCI students and older adults working on the topic of technology-supported local caring communities.

Cross Community Learning plays a major role in IT design. Concepts such as participatory and co-design focus on interdisciplinary learning relationships with building up spaces where mutual learning at eye level is possible. With the entry of digital tools into all areas of life and also to user groups that until recently were not considered as target groups of IT production, e.g. people with disabilities, children or older people, a further development of a systematic understanding of learning, IT design processes, and the social space both is situated in is necessary.

„When the data is created from you, about you by you, then you should have the right to access and use that data above and beyond anybody else“ – Annika Wolff

Annika Wolff spoke on this topic with the focus ‚Facilitating Data Sense-Making for Civic Empowerment‘. Behind the strong title lies a local, situational understanding of data and data practices, which are subjectively produced by human and non-human actors against the background of their respective contexts and result in further interaction between humans, machines and the data stream. Emotion emerges from this interaction. Data therefore has a collaborative nature that connects actors. Whether this connection is based on consensus cannot be assumed, however, according to Wolff.

Consensus presupposes a common understanding. Data must therefore be universally interpreted in the same way, but this often proves difficult in practice, because within our data there is often messiness in form of errors and lack of information.

Within the common understanding process, further errors can also occur. Background: Marginalized and/or vulnerable target groups in particular, who have been largely ignored in the development of digital technologies to date, struggle in their understanding of data and/or are underrepresented in existing data, e.g. due to a lack of accessibility to technology or integration into the design process.

Starting from the problem, collecting and processing data to achieve a solution, whether the problem is of a societal, target group-oriented or individual nature, is therefore always a challenge. However, it is even more difficult to localize the problem from the existing data set and then generate an adequate solution. The collection of similar data, which is referred to as the ‚personal data collection principle‘, can support this. This principle opens a door that makes especially those heard who usually cannot speak for themselves. Nevertheless, data availability does not automatically mean data justice, an individual or target group-based problem that remains as a societal challenge.

How can we adequately address this dilemma?

Against this background, Wolff describes local-oriented and human-centred methods that promote empathy and empowerment of the civilian population by incorporating artistic aspects and focusing on the promotion of media competence via data. The so-called ‚data drama‘, which she cites as an example, allows for immersive data exploration through data games, where the dramaturgical framework functions as an alternative way of making sense of data, which is generated in interactive, collaborative contexts.

Open questions are those concerning the adequate communication of the benefits of data to various target groups and, in this respect, those concerning the first steps towards joint data exploration.

„If we don’t have good data we can’t go on with Cross Cultural Learning.” – Jennifer Rode

Cross Cultural Learning is an important aspect of Cross Community Learning. Jennifer Rode deepened this basis of shared knowledge culture in her contribution ‚Taking on Femininities+ for Cross Cultural Learning‘, in which she focused on the participation of minorities in the STEM sciences. While the call so far has been for more women and thus equality within the commonly held binary system, Rode argues for moving beyond this familiar and limited framework, which so far excludes the current 10% of people who identify as non-cisgender.

The social construction of systematic binary education within our societies not only creates a very simplified image of gender, but is at the same time, according to Rode, tied to a technical identity. However, societies differ culturally in this respect. A transformation towards Cross Cultural Learning is thus also linked to a transformation within the framework of social inclusion; value systems must be opened up and expanded to include those that have so far remained unconsidered. Only such an integration and recognition of multiple perspectives will ultimately enable the inclusion of equal values in technical design, which shapes lives worldwide. Such equality of diverse genders and cultural backgrounds thus reduces the inherent bias in our current technologies and thus data practices and data streams and therefore reduces the subliminal racism fed into technical artefacts that we have to deal with today in a self-critical and self-reflective way against the background of Western, masculine and white privilege.

On a social level, such an opening through inclusion gives authenticity to those marginalized groups who have so far been subjected to the mantra of gender-oriented performance, especially in professional contexts. It also creates equal participation at all levels and enables the use of one’s own voice in private and public spaces. Ultimately, everyone benefits from this, including the male gender, according to Rode, which until now, exposed to toxic masculinity, has also been forced to perform.

The guiding principle for further research and practical implementation of ethically justifiable IT design is therefore:

„Embrace femininities+ and intersecionality as starting points for design and theoretical framework to examine the bi-directional shaping of society and technology.“ – Jennifer Rode

A practical example of cross-cultural learning against the background of such a transformative process was given by Verena Fuchsberger with the title ‚Making, Makers and Makerspaces: Access and Other Barriers‘. By integrating female engagement within this field, the previous trend of primarily male presence in makerspaces is breached. Another important factor is education. Makerspaces can be understood as educational spaces that, contrary to their current nature, should not be homogeneous, but a safe, social space for people with different levels of education, where everyone can be both teacher and student. However, according to Fuchsberger, this wishful thinking is sometimes accompanied by barriers. They include the overlapping of subject and skill areas, consisting of arts and crafts and engineering. As Jennifer Rode pointed out, our technical identity is tied to our gender identity. In practice, this creates a perception for many people that they may not be technically skilled, or even a fear of failure. This attribution of competence or lack thereof can further lead to an inner resistance that prevents participation in order to preserve one’s autonomy. During participation, such a feeling can lead to a sense of being overwhelmed. The negative influence of one’s own socialised peer group, which warns of such barriers and further negative experiences or emotions in advance, has also been identified.

What is needed are new role models. According to Fuchsberger, pioneers within male-dominated spheres, such as in the makerspace, are necessary to act as door openers and lower fears and barriers. Already existing or increasing diversity must be made more visible against this background.

„One day, maybe, we will ask: ‚What was the gender gap?‘ and ‚How was it solved?‘ and ‚Why did it even exist?” – Verena Fuchsberger

On the way to such a visionary future of equality, Cross Community Learning against the background of a common understanding of data, in new safe social spaces, represents the cornerstone, as does its theoretical foundation, which is based on a change in values within science, which, as it were, carries such a change back into society via technical design and helps to reshape socialisation.

Gerhard Fischer took a closer look at the future in his keynote ‚The Future of Learning and Digital Media. Exploiting and Supporting the Synergy between Renaissance Scholars and Renaissance Communities‘.

„Collaboration is key as the power of the unaided individual human mind is limited.“ – Gerhard Fischer

This orientation in the here and now is accompanied by many questions, e.g.: What role can technology play in this? What can meaningful knowledge management look like? or How do we move from a culture of ‚having to learn‘ to one of ‚wanting to learn‘?

With reference to antiquity, Fischer points to the importance of external forms of documents that have always carried the knowledge of mankind to ensure its collective survival. Modernity and its technical designs, however, have led us into a dilemma. Thus, in addition to truths, negative effects are now hidden in our knowledge tools, such as information overload or filter bubbles. In addition, we are struggling today with Wicked Problems of global proportions that are bringing our established systems and familiar structures to their knees and making change inevitable. Topics such as technology development, participation and learning are coming to the fore. Keywords here are Cross Community Learning, Cross Cultural Learning, Situated Learning and Life Long Learning. Learning is key. Individual attributes as well as those of the respective context come to the forefront of research and development, roles of teacher and learner become increasingly blurred, and rightly so, because:

„Intelligence is distributed across minds, across cultures.“ – Gerhard Fischer

Traditional learning systems, into which technology is incorporated, will compete in the future with such – still visionary – co-evolution of people, knowledge and technology, which form a new, shared learning space. Covid-19 can be interpreted as a push into such a development, Fischer says. How do we design learning spaces under these conditions and henceforth to turn the compulsion for transformation into one of added value?

Fischer speaks here of Renaissance Scholars or Renaissance Communities because we are in a time of transition. Just as the Renaissance, which describes the transition from the Middle Ages to Modern times, was characterised by a revival of cultural achievements of antiquity, we need to revive cultural achievements to find our way back to the collective intelligence, which is composed of diverse forms of knowledge and culture. The very impossibility of a perpetual articulation of specific problems makes inclusion of affected persons or groups in participatory learning processes indispensable.

„None of us is as smart as all of us.“ – Gerhard Fischer

In addition to cultural knowledge, other different forms of knowledge are waiting on the path of transformation towards a collective knowledge structure, e.g. professional knowledge, technical skills or media competence. A model for the Renaissance of the 21st century therefore aims at sharing knowledge and learning from each other, which is only possible in open, participatory learning spaces. This is where common ground must be established. A legitimate question here is:

How do we move from a competitive society to one that is open-minded and shares its knowledge openly?

Fischer refers to the so-called Fish-Scale Model (Campbell 1969): „collective comprehensiveness through overlapping patterns of unique narrowness“.

A first step is to promote interaction and communication. This is especially possible at interdisciplinary interfaces. Creating common ground through shared knowledge is fundamental here, with technological artefacts – boundary objects – playing an essential role in supporting the process of externalising ideas that support shared understanding across spatial, temporal, conceptual and technical gaps. The intertwining of Renaissance Scholars and Renaissance Communities is seminal against this backdrop, with individuality and its imprint in particular making a difference.

Who designs time and its use?

On the way to the transformation described above, so-called meta-design is becoming more and more important: participatory design for designers, not exclusively for users, or the latter should also become a designer representative of the entire user group and help shape future social and digital practices.

A second model as a possible anchor point for the development towards Renaissance Communities, to which Fischer refers, is SER (Seeding, Evolutionary Growth, Reseeding). Here, the user as designer is assigned a leading role, especially in the second phase, of Evolutionary Growth, as he or she takes the lead completely independently of the professional designer, i.e. he or she is empowered to drive development autonomously during this time.

The empowerment for this comes from the development of supporting socio-technical environments, which are shaped by Communities of Learning (CoL) that come together against the background of common interests. In order to pool knowledge on a large scale, data-oriented approaches make sense despite all the challenges, e.g. information overload or data manipulation. They create common ground while taking validity into account. But caution is advised:

„Often we value what we measure. Instead, we should measure what we value.“ – Gerhard Fischer

Research questions that follow on from this are: What data do we have? What problems can this data address? What problems do learners have? and What data are needed?

In conclusion, it can be said that against this backdrop, at the vertex between downfall and new beginning – in the dawn of a new Renaissance, modernity has a new mantra to internalize:

„The future is not out there to be discovered. The future has to be invented and designed. – Gerhard Fischer

What is and remains important is the question of who. Who is currently shaping the future and who should?

Further information on our referees:

Annika Wolff

Background & research interests:
Annika Wolff is an assistant professor whose research is in the newly emerging field of human-data interaction, at the intersection between complex data, machine and human learning. Her research focuses on engaging people with data, such as from smart cities, and in supporting non-experts in designing products and services that use data. She has expertise in using inquiry-based methods and co-creation, in developing applications of data science and in the use of tangibles, creativity and games to support learning. She has previously led work in developing and piloting new methods for teaching data literacy skills in UK primary and secondary schools and in understanding how open data can be utilized in education. She has many years of experience working within UK and European funded projects, combining applications of data science to human understanding. She has published in a number of international journals and is an active member of research communities related to community-based innovation and HCI.

More information:
https://research.lut.fi/converis/portal/detail/Person/9873617?auxfun=&lang=en_GB

Jennifer Rode

The speaker about her research interests:
“My research lies in the areas of Human-Computer Interaction and Ubiquitous Computing. I am deeply committed to understanding why technology is so empowering to some and for others it is the foundation for exclusion. My work examines the values of technological utopianism and how those values influence the user-centered design process. Ultimately, I look at how, and under what circumstances, individuals choose to use technology. Consequently, my work explores areas in which the experiences of both enthusiasts and technophobes overlap. This makes technology in the home especially relevant as it is a primary environment in which we socialize children in socially approved attitudes towards technology. My research looks at how gender is key to the ways technology is both perceived and used. Further, it looks reflectively at the design process to see how our biases and practices shape the artifacts we design.”

More information:
http://www.jenniferarode.com/Jennifer_Ann_Rode_CV2017.pdf

Verena Fuchsberger

Background & research interests:
Verena Fuchsberger is Postdoc at the Center for Human-Computer Interaction, University of Salzburg. She has completed her Master’s Degree in Educational Sciences and Psychology at the University of Innsbruck and finished her PhD in HCI at the University of Salzburg in 2015. In her research, Verena focuses on the agency of human and non-human actors in HCI and interaction design, i.e., the relations and associations between individuals (users, but also non-users, designers), materials and digital-physical artefacts. In particular, she is interested in the hybrid materiality of interactions as it occurs in different domains, contexts, and situations.

More information:
https://hci.sbg.ac.at/person/fuchsberger/

Gerhard Fischer

The speaker about himself:
“I am a Professor Adjunct and Professor Emeritus of in the Department of Computer Science, a Fellow of the Institute of Cognitive Science, and the Director of the Center for Lifelong Learning and Design (L3D) at the University of Colorado, Boulder. I am a member of the Computer Human Interaction Academy (CHI; 2007), a Fellow of the Association for Computing Machinery (ACM; 2009), and a recipient of the RIGO Award of ACM-SIGDOC (2012). In 2015, I was awarded an honorary doctorate from the University of Gothenburg, Sweden. My research interests include: new conceptual frameworks and new media for learning, working, and collaborating, human-centered computing, and design. My recent work is centered on quality of life in the digital age, social creativity, meta-design, cultures of participation, design trade-offs, and rich landscapes for learning (including MOOCs). You can find out more about my work by clicking on the respective menu entries for: Publications, Reports and Presentations, Recent Courses, Major Research Projects, and Ph.D. Graduates.”

More information:
https://l3d.cs.colorado.edu/wordpress/people/home-folders/gerhard-fischers-home-page/

>click here to enlarge image<

Graphic recording of Annika Wolff’s talk on „Facilitating Data Sense-Making for Civic Empowerment“, as part of panel III „Cross-Community Learning“ on Oct 28, 2021. Graphic recording by Coline Robin https://coline.graphics/

>click here to enlarge image<

Graphic recording of Jennifer Rode’s talk on „Taking on Femininities+ for Cross Cultural Learning“, as part of panel III „Cross-Community Learning“ on Oct 28, 2021. Graphic recording by Coline Robin https://coline.graphics/

>click here to enlarge image<

Graphic recording of Verena Fuchsberger’s talk on „Making, Makers, and Makerspaces: Access and Other Barriers“ as part of panel III „Cross-Community Learning“ on Oct 28, 2021. Graphic recording by Coline Robin https://coline.graphics/

>click here to enlarge image<

Graphic recording of Gerhard Fischer’s keynote on „The Future of Learning and Digital Media. Exploiting and Supporting the Synergy between Renaissance Scholars and Renaissance Communities“, as part of panel III „Cross-Community Learning“ on Oct 28, 2021.

Graphic recording by Coline Robin
https://coline.graphics/

One of the key questions for the Panel on Human-Machine-Learning is not simply how machine learning (ML) ‘works’, but how it enrols humans in new kinds of work. The talks in this session touch on various parts of this relationship, from the ‘looping’ of humans into ML processes, to the role of ML tutorials in teaching software developers how to use ML methods. One key concept I perhaps want to draw attention to here is the notion of ‘experience’, and how experience is often a critical feature of human-machine-learning relations.

Heuer and Jarke’s abstract, for instance, mentions how ML tutorials are presented as both ‘universally applicable’ (meant for any kind of use, or deployment in any domain) as well as implementable ‘without special expertise’. That is, without any specific education in how ML algorithms ‘work’ or how data (its variety, volume, granularity etc.) is critical to how successfully ML can be deployed. But more importantly, I think, in how such tutorials frame ML as readily, and easily, ‘dropped into’ any existing workflows a software developer might have. Here, ML deliberately appears as an ‘off-the-shelf’, ‘ready-to-use’ thing without any need – as Heuer and Jarke’s analysis of ML tutorials suggests – to know how they work. In this, tutorials function to persuade practitioners that no prior experience is necessary.

Experiencing the joy of the Waymo Driver

In the case of the autonomous vehicle company Waymo, experience is used somewhat differently. Here, experience is a goal in itself and specifically, a mission to build the ‘World’s Most Experienced Driver™’ (Waymo 2021). Here, experience takes multiple forms. Firstly, experience is rendered in metric form in miles driven, and time driving: over 20 billion, both real world and simulated, over a decade of development. Secondly, experience is rendered in situational form, as ‘challenging, educational, and groundbreaking’ (Waymo 2021, n.p.) miles accrue in a continual, ongoing effort to build the most experienced autonomous vehicle. Thirdly, in affective form, as a smooth, safe, trouble-free ‘user experience’ (beyond aesthetics or design) for those inside and outside the vehicle. Understanding how such vehicles undertake rote driving tasks, whether learning to park or merge lanes, routinely discussed when talking about ML in these instances, only captures a small part of what is going on here. Much work, it seems, is being dedicated to cultivating experiences in multiple ways.

The notion of experience also invites two further thoughts. Firstly, the importance of temporality, seen as critical to accruing or amassing experience. Training data used in ML applications can perhaps be seen as injecting near-immediate experience into an ML process, but does this short-circuiting of the relationship between experience and time accrued also result in a greater lack of trust? Secondly, picking up on the discussion above, the analytical advantages of distinguishing between ML practices or operations and ML experiences. In this, do ML processes have character or style derived from the entanglement of (training) data, humans-in-the-loop, weightings and parameters, and to what extent are we able to make sense of how an ML-based application changes character, subtly or not, through the experience it amasses? If our experiences shape us, how do they shape them, with or without our help?

References

Waymo (2021). We’re building the world’s most experienced driver™. Available: https://waymo.com/ (accessed 22 October 2021).

Sam Hind, research associate in SFB1187 Media of Cooperation in project A03, ‘Navigation in Online/Offline Spaces’, sam.hind@uni-siegen.de

Machine Learning (ML) has become a key component of contemporary information systems. Unlike prior information systems explicitly programmed in formal languages, ML systems learn from data. We discuss why existing approaches are not useful for the critical analysis of socio-technical systems based on ML. We argue that algorithms are not the central issue. Since a mutually shared understanding of Machine Learning is still missing, we examined how ML is understood in practice by analysing how ML is framed in tutorials. As sources of informal learning frequently used by professional software developers, ML tutorials enabled us to gain a broad overview of the different definitions of ML and the types and algorithms of ML recognised by practitioners. Our systematic analysis of 41 ML tutorials revealed which applications of ML are used as examples and which parts of ML are explained how. Our analysis identifies canonical examples of ML as well as important misconceptions and problematic framings: While algorithms do play a marginal role, the importance of data is vastly understated. Most importantly, we find that ML is presented as universally applicable and as something that can be implemented without special expertise. We argue that attention should be paid to how ML is framed and executed as a new paradigm of computing. We conceptualise ML as part of a complex socio-technical system and extend on prior work in critical scholarship by shifting the analytical focus from algorithms to data.

Jarke, J. & Macgilchrist, F. (2021). Dashboard stories: How the narratives told by predictive analytics reconfigure roles, risk and sociality in education. /Big Data & Society/. Open Access. → full paper: https://journals.sagepub.com/doi/full/10.1177/20539517211025561

Heuer, H., Jarke, J. & Breiter, A. (2021). Machine Learning in Tutorials: Universal Applicability, Underinformed Application, and Other Misconceptions. /Big Data & Society/. Open Access. → full paper: https://journals.sagepub.com/doi/full/10.1177/20539517211017593

Mercedes Bunz: Error Is No Exception. On Machine Learning, Errors, and Phil Agre’s Critical Digital Practice

Mercedes Bunz took up the conference description of machine learning systems as models trained by an “ensemble of data, algorithms and practices of modeling and evaluation” and called its elimination of the learning subject into question. While humans learn to recognize objects relatively quickly through imitation, algorithms require large amounts of examples (data) and training to properly classify objects (statistically). Image processing programs, on which Bunz focused, also learn by looking at the smallest element of an image such as shades and edges; a process that she described as “calculation of meaning”.

Receptions of current applications of AI regard them as artificial copies of human intelligence. With Simondon, Bunz described these copies as “machines in the service of man” (Simondon 1958: 17): a technology without agency. Bunz held this perspective of a pure, smoothly functioning and fixable technology against a variety of errors in image recognition which are – as she stressed – inherent to machine learning systems.

The prevalent assumption of an always rational and error-free technique becomes obvious in the reaction of the developers towards an error by Alpha Go

Whereas the mis-categorization of images and meaning was long seen as man-made error, so-called adversarials were created to trick the program’s perception with input samples, thereby demonstrating the fallibility of image recognition algorithms, i.e., their misinterpretation of meaning. By adding a layer of noise, adversarials lever out the algorithm’s focus on the smallest element of images. Misinterpretations of “natural” images, of photos of real-world objects, have also been found, and thus support the thesis of errors as inherent in technology: “Adversarial examples are not bugs, they are features” (Ilyas et al. 2019). [1]

examples of mis-categorization

The technicality of systems becomes obvious in these moments of non-functioning. What for Heidegger is the knowledge object and for Rheinberger the epistemic thing, Bunz grasped with Virilio, who understands accidents as starting points for a techno-analysis of the substance of knowledge. Since the error is recurrent and cannot be fixed, the black box remains open, prompting Bunz to call for an ongoing collaboration of machine learning systems and human beings. She followed Agre’s call for “collectively produced technology” in which scholars of different disciplines and communities work version-based together on shaping algorithmic practices. With their need of a constant practice of adjustment and a focus on the actual situation, machine learning systems thus become media of cooperation.

[1] Ilyas, Andrew, Shibani Santurkar, Dimitris Tsipras, Logan Engstrom, Brandon Tran, Aleksander Madry 2019. “Adversarial examples are not bugs, they are features.” arXiv:1905.02175 [stat.ML].

Paola Tubaro: Learners in the Loop

Agre frames collectively produced technology positively, as a critical exploration of the situation by engineers, media and technology scholars and the involved communities. Paola Turbaro attends to the realities of contemporary AI work, in which large numbers of workers collectively train algorithms without the ability to critically engage with their work, sometimes even without knowing what they are working on.

Her inquiries into the role of human subjects within contemporary forms of AI production shift the focus from the ‘human in the loop’ to both human and non-human ‘learners in the loop’, emphasizing the labor involved in creating and maintaining the datasets underpinning the machine learning algorithms used in Intelligent Personal Assistants (IPA) such as Alexa or Siri. Drawing upon an extensive corpus of questionnaires and interviews, she identified three areas of interest involved in the ‘teaching of the machine’: 1) data creation carried out by speakers, 2) data annotation carried out by writers and 3) the verification of results carried out by listeners.

Tubaro, P., Casilli, A. A., & Coville, M. (2020). The trainer, the verifier, the imitator: Three ways in which human platform workers support artificial intelligence. Big Data & Society. https://doi.org/10.1177/2053951720919776, https://journals.sagepub.com/doi/full/10.1177/2053951720919776

Hendrik Heuer and Juliane Jarke: From Algorithms to Data

While Tubaro pointed to the invisibility of human labour when dealing with Machine Learning, this opaqueness is also evident in the process of programming. As Heuer and Jarke showed in their study, only 1% of software developers have formal expertise in working with AI and Machine Learning. Most commonly knowledge about such applications is acquired through online communities and in the form of self-education. Especially online tutorials play a crucial role in the acquisition of knowledge.

Python Machine Learning Tutorial

Most of these tutorials frame Machine Learning as a tool that is universally applicable and implementable without special expertise, e.g. for spam detection, house price prediction or self-driving cars. Within this focus on application, not only the actual functioning of the algorithms recedes into the background, but also the crucial role of the data fed in. Heuer and Janke saw an imminent risk in this form of black boxing. In order to recognise the social impact of Machine Learning and AI, it is necessary to break the narrow focus on algorithms and their application and to understand the data practices behind them.

Heuer, H. 2020. Users & Machine Learning-Based Curation Systems (Doctoral dissertation, Universität Bremen).

Heuer, H., Jarke, J., & Breiter, A. (2021). Machine learning in tutorials – Universal applicability, underinformed application, and other misconceptions. Big Data & Society. https://doi.org/10.1177/20539517211017593, https://journals.sagepub.com/doi/full/10.1177/20539517211017593

Sam Hind: ‘What Ifs?’ Path Predictions and Counterfactuals

According to Sam Hind, the death of Elaine Herzberg – an incident in which an autonomous Uber vehicle killed a pedestrian after misclassifying her and misjudging her travel path multiple times in a short time span – and the ensuing reputational damage to the whole branche prompted companies to restage and resituate the safety of autonomous vehicles.

Taking a closer look at Google’s Waymo, who now euphemistically list their vehicle’s “contact events”, i.e. accidents of varying severity, Hind unfolded the concept of deferred decisions: decisions which are never final and only act as data points for future decisions. These come into play when the car encounters a situation in which the human safety driver has to take over: the moment before the ‘crash that never came to be’ provides the basis for a multitude of counterfactuals, i.e. simulated crashes aimed at improving the vehicle’s grasp of the situation.  Here the error – which Mercedes Bunz showed to be essential to machine learning – is utilized in a slightly different way. While it is not allowed to happen in the physical space, it gets transposed to a simulational crash test site which can then be explored without harming other road users. Hind dubs these seemingly innocuous moments which yet prompt the production of counterfactuals – and eventually learning progress in the ML model – as ‘sliding door moments’, highlighting these simulational ‘what-if’ approaches as yet another aspect of machine learning that remains to be rendered visible by scholarly engagement.

>click here to enlarge image<

Graphic recording of Paola Tubaro’s talk on „Learners in the Loop“, as part of panel IV „Human-Machine-Learning“ on Oct 29, 2021. Graphic recording by Coline Robin https://coline.graphics/

>click here to enlarge image<

Graphic recording of the talk by Hendrik Heuer and Juliane on „From Algorithms to Data“, as part of panel IV „Human-Machine-Learning“ on Oct 29, 2021. Graphic recording by Coline Robin https://coline.graphics/

>click here to enlarge image<

Graphic recording of Sam Hind’s talk on „‘What Ifs?’ Path Predictions and Counterfactuals“, as part of panel IV „Human-Machine-Learning“ on Oct 29, 2021. Graphic recording by Coline Robin https://coline.graphics/

>click here to enlarge image<

Graphic recording of Mercedes Bunz’s keynote on „Error Is No Exception. On Machine Learning, Errors, and Phil Agre’s Critical Digital Practice“, as part of panel IV „Human-Machine-Learning“ on Oct 29, 2021.

Graphic recording by Coline Robin
https://coline.graphics/