Fishbowl 3.1: When “open” isn’t: open-washing and the politics of openness in the age of Generative AI. Anna-Maria Sichani, Marina Markellou, Douglas McCarthy
- Started with a mentimeter – cool idcea that asked “what does open mean to you?”
- Paul Keller and ALek Tarkowski, The Paradox of Open
- Digital Commons EDIC launches to advance Europe’s technological sovereignty
- Publishing cultural heritage in the age of AI
- Beyond AI and Copyright
- Open Source AI Definition:
Open Source has demonstrated that massive benefits accrue to everyone after removing the barriers to learning, using, sharing and improving software systems. These benefits are the result of using licenses that adhere to the Open Source Definition. For AI, society needs at least the same essential freedoms of Open Source to enable AI developers, deployers and end users to enjoy those same benefits: autonomy, transparency, frictionless reuse and collaborative improvement.
- Mozilla Data Collective
- CC Signals AI within Creative Commons
- https://www.dataspace-culturalheritage.eu/en
Research & Education
Claudia Montanaro – What ‘Meaning’ Means in LLMs’ Research: An Interdisciplinary Conceptual Map
- LLMs outputs meaningful ideas
- Can LLMs make sense? Can they represent meaning? Noam Chomsky interview and his answer is “It’s like asking whether submarines swim”
- Hemran Cappelen – it can make language make sense
- Difficulty in borrowing terms. Key terms are borrowed from human cognition and are ill-defined
- Method: dataset of academic articles , used ATLAS.ti
- Identified six clusters
- Three main themes: meaning is measurable, meaning as emergent property, meaning as something to be understood
- Meaning and understanding in LLMs is
- The call for scholars from a rnage of fields ot understna
Q: Can you explain the terms a litter further? unpack the network map how does that relate to the table?
Maede Mirsonbol – Learning with GenAI Images: Supporting Higher Education Students’ Reflection on Inclusive Education
- AI literacy in in higher education, we need different frameworks to introduce to students
- Relational Pedagogy – Concept and Contest Teacher <-> Learner <-> Content
- Where to place AI?
- Leaning on UNESCO guideline on AI and Education
- Student <-> Teacher <-> Text <-> AI, Student learning model
- Two case studies, Lancaster University, England, and University of Tartu, Esrtonia
- Groups learning using a semiosis-based method to learn with images
- 5 Steps of inclusive as diversity education: exploration (what is AI education?), co-creation (groups prompting),
- Then they create images of inclusivity (race, gender, access/communication)
Q: what is the take up of this research like in higher ed? Did participants understand diversity within the four categories demographic, communication, species, nature/material/place?
Berk Alkoç – Generative AI in Design Bootcamps: A Critical Inquiry into Pedagogy, Dependency, and Creative Practice
- Figma and Make Design, prompt with an idea and it will make an app, borrowing the Apple version/platform/tool
- Bootcamp and education – Careerfoundary UX Design bootcamp
- Discourse analysis of marketing and and advertising of a number (10?) UX design bootcamp – a determinism built in of ‘if you don’t know AI, you won’t get a job”
- How to create a persona in 5 minutes – this demonstrates a ‘frictionless’ process, where the process is more important than the outcome
- Are they training designers or AI operators?
- Reference to Hooked
- Design fixation – pre-existing solution and restricts creative thinking
- PLatform deactivation very similar to the API reliability of platforms
- Should we go backwards and ignore the algorithm? (can we though? we’ve ‘drunk the Koolaide’)
Q: what does the industry want/what does the environment look like? Is this quick turnaround what is needed as we build design expertise? what does a co-designer model look like?
Panel 5.8: Creative Work
Nirvi Maru, Vivian H. H. Chen – A Human-AI collaboration approach to creative workflow
- A GenAI inflection point – democratisation vs. diminishment (erosion of human agency, homogenization), to be understood not as a binary
- Literature approaches this as a discrete problem to be solved
- scoping review – 216 papers
- Conceptual fragmentation – collaboration, teaming, autonomy, etc.
- Trust-calibration
- Erosion of human agency – skill degradation,
- task collaboration misalignment – poor fit between AI capabilities and workflow
- Reframing Human-AI Creative Collaboration (HAIC): interdependence, hidden costs, navigation tensions
- Agency Automation Tension – tension needs to be managed not solved
- Principle 1 – ROle Design as Creative Act: we negotiate the ‘role’ within HAIC
- Principle 2 & 3: Transparency and Human Experience. As a creative uses more AI, they trade off the critical aspects of design. Human experience should be centred.
Yunus Emre Öztaş – How Creative Workers Make Do with GenAI in Visual Media Production
- Critical creative labour studies, mostly in cultural and creative industries – high value in autonomy within an environment of augmentation-replacement binary
- A spectrum of GenAI use model (in iterative development): Human Creative Agency Dominant (task related use) <-> Machine agency dominated (task-agnostic use)
- Modes of use are embedded in agency
Tolulope Oke, Robert Prey, Femke de Rijk – Beyond the Global North: Generative AI and the Future of Musicians’ Work in Nigeria
- Based in Nigeria Music Market, growing rapidly (three times larger than the rest of the world’s music industry)
- There is a talented yet unstructured market (copyright and infrastructures are lacking)
- Futures of the industry which is usually dominatd by the Global North, where GenAI becomes a crucial infrastructure
- Africanfuturism – visions of the future, not concerned with what could have been but what is possible/what is the future
- Decolonial AI: extends beyond data colonialism and into all material aspects of data generation (check this)
- Future reference about Phillips Olajide, First African trained AI music generator
- Korin AI – social-technical artefact that reflects and shapes how Nigerians imagine and navigate AI Futures
- Jamai Fabuyi and legal contracts in Africa – interview here
- Conclusions: opportunity to restructure global industries. It’s not about if it’s changing it’s about how it’s changing.
Closing Panel with invited speakers – Pei Sze Chow, Maximilian Schich, Naureen Mahmood
Pei Sze Chow
- Specificity & pluralism across the variety of talks across the past few days. Contractual obligations and the sorts of creative work and their ecologies.
- Skills. De-skilling. Up-skilling. the ways in which creatives are adopting new skills, retool a skill-set, etc.
- (my thought – can we stop with the ‘democratisation’ as waves of media technologies emerge. We saw this in Web 2.0, social media, and in some respect through platformisation, and we know this is never the case)
Naureen Mahmood
- Meshcapade (swing this to Rangi and see what he thinks)
- Job displacement – everytime a technology revolution, it’s never a loss but a change. “how we let that happen is up to us”
- Fear around AI and how others are using it (i.e. Governments especially). This community has a lot to offer in terms of understanding what it actually is and the kinds of applications that are possible.
Maximilian Schich








