Kurs Critical AI

Prof. Dr. Artemi-Maria Gioti

Application for ditact_women’s IT Summer Studies 2025

Course Title:     Critical AI

Course Level:    Interdisciplinary Course

Course Duration:    Two-Day Course (14 EH)

Instructor:    Prof. Dr. Artemi-Maria Gioti

1. Course Description

This two-day workshop will explore critical perspectives on Artificial Intelligence (AI) and Data Science from the arts and humanities and examine how these perspectives can inform technical development in these fields. AI systems and data infrastructures are often assumed to be neutral and universal, yet they embed cultural assumptions, historical biases, and epistemological frameworks that shape their outputs and societal and cultural impact. By engaging with critical theory, artistic research and interdisciplinary discussions, participants will analyze the ways in which AI constructs knowledge and how alternative perspectives can contribute to a more reflexive AI design.

Key questions addressed in the workshop include:

•    What assumptions and values are embedded in AI systems and data?

•    What do these assumptions reveal about the epistemology of AI?

•    What can artistic encounters with AI reveal about the inner workings of Machine Learning (ML) algorithms and data?

•    And how can artistic research and humanities-based critiques inform AI development?

2. Course Format

The workshop will blend theoretical inquiry, practical experiments, and creative ideation, guiding participants to critically examine AI systems and their implications through interdisciplinary lenses.

2.1. Discussion of selected readings 

The workshop will be structured around guided discussions led by the instructor, focusing on selected readings from key critical AI literature. Participants will engage in close readings, discussions, and reflections, bridging insights from humanities and the arts with technical concerns in AI development.Core readings will include:

•    Kate Crawford & Trevor Paglen, Excavating AI: the politics of images in machine learning training sets – A seminal analysis examining the work that image datasets – and their labels – do and the values behind the taxonomies they construct.

•    Donna Haraway, Situated Knowledges: The Science Question in Feminism and the Privilege of Partial Perspective – A foundational text in feminist epistemology, challenging the notion of scientific objectivity and advocating for embodied, situated perspectives.

•    Catherine D’Ignazio & Lauren Klein, Data Feminism – A critical examination of power dynamics in data science, highlighting issues of bias, inequality and social justice.

•    Philip Agre, Toward a Critical Technical Practice – An argument for embedding reflexivity and critique into technological design, bridging engineering and humanities perspectives.

•    Safiya Umoja Noble, Algorithms of Oppression – A deep dive into racial and gender biases in search engine algorithms.

•    Wendy Chun, Discriminating Data – A study on how modern data practices reinforce historical inequalities.

•    Nick Seaver, Algorithmic Recommendations and Synaptic Functions – A discussion of how recommendation systems function as cultural intermediaries, shaping users’ preferences and experiences.

•    Artemi-Maria Gioti et al., Composing the Assemblage: Probing Aesthetic and Technical Dimensions of Artistic Creation with Machine Learning – An artistic research perspective focusing on a deconstructive critique of data as materially contingent and imbued with aesthetic values.

•    Mark Andrews, The Devil in the Data: Machine Learning & the Theory-Free Ideal – A critique of the assumption that data-driven approaches are free from theoretical bias and subjectivity.

2.2. Practical Experimentation 

Participants will engage in hands-on experiments in data collection and machine learning model training. This practical component will focus on the ontology of data, exploring popular narratives surrounding data and how these might contradict current ML and data practices.

2.3. Creative Ideation Sessions 

In these sessions, participants will engage in forward-thinking exercises that challenge conventional AI development approaches:

2.3.1. Reverse Engineering: Participants will critically analyze existing AI systems focusing on the wider cultural, historical, political and economical contexts within which they are embedded and operate and the values crystallized in them.

2.3.2. Design Fiction: Participants will create speculative scenarios to imagine new AI systems grounded in alternative epistemological, cultural and other assumptions.

3. Target Audience

This workshop is designed for IT practitioners, students and researchers interested in integrating critical perspectives into AI and Data Science. It is accessible to participants from a variety of disciplines and backgrounds. No prior knowledge is required for participation.