Dates: 18–20 February 2026
Format: on-site
ERUA and New Bulgarian University (NBU) are announcing an intensive course “Emotional noise in generative art and AI” that will take place at NBU, between the 18th and the 20th of February 2026. The course will consist of daily meetings from 10:00 to 16:00, group work, practical sessions and discussions.
This 3-day intensive course explores the emotional and subjective dimensions of human experience in the context of generative AI. Through a combination of short lectures, group exercises, tool experimentation, and critical reflection, students will be guided to build a small, emotionally-driven visual dataset and use generative AI models (such as DALL·E, Midjourney, or Stable Diffusion) to interpret non-objective, affective prompts. Teaching methods include: experiential learning, critical discussion, guided tool use, and peer feedback.
Programme:
Day 1: Emotional Noise as Missing Data
Introduction to the problem of emotional invisibility in AI training data. Students reflect on the limits of objectivity in datasets and identify personal images or feelings that are hard to describe in words.
Day 2: Subjective Dataset Creation
Students create a micro-dataset (5–10 images) with both objective and subjective emotional labels. They critically compare automatic descriptions (e.g., from CLIP Interrogator) with personal interpretations.
Day 3: Generative Interpretation & Reflection
Students generate images using AI tools based on their emotional labels. They compare results to their original inputs and write short reflective texts.
The course will conclude with a group discussion on the limitations and potentials of AI as a medium for subjective understanding.
Learning outcomes:
Students will acquire:
- Ability to critically assess emotional subjectivity in AI systems;
- Skills in working with generative AI tools (e.g., prompt crafting, image evaluation);
- Reflection on data ethics and emotional bias;
- Creation of a subjective, human-centered micro dataset.
Assessment:
The course grading will be based on continuous assessment through participation, creation of a visual dataset, written reflection, and final group presentation.
Upon successful completion, the course will award 2 ECTS.
Who can apply:
We are looking for 10 BA and MA students who are enrolled in ERUA member Universities and who have a strong interest in digital art, visual culture, or AI and society, and a good command of the English language. No advanced technical skills are required.
Important: Students must have access to at least one generative AI platform (e.g., ChatGPT Plus, Midjourney subscription, or access to Stable Diffusion platforms) in order to fully participate in the practical exercises.
Place: New Bulgarian University, Sofia, Bulgaria.
Application deadline: 20th of December 2025
How to apply:
To apply please fill in the online form and send a motivation letter in English together with a portfolio via e-mail to the course coordinator, Dr. Ivaylo Saraliyski, at isaraliyski@nbu.bg.
Selection will be based on motivation, interest in the topic, and diversity of background.
Funding:
ERUA students can receive travel and accommodation funding from their home universities. Please contact your local ERUA coordinators.
