Concept
NOT PROMPT — media examining the impact of generative artificial intelligence on contemporary visual culture. It investigates how neural networks are changing design, photography, fashion, and art, as well as how people relate to images.The core idea behind the media is that artificial intelligence is becoming not merely a tool for creating visual content, but an independent force shaping the visual environment. Image generation raises new questions about authorship, originality, aesthetics, and the boundaries of human creativity.The primary value of NOT PROMPT is the critical contemplation of AI-generated visual content. The media does not teach prompt writing or how to use neural networks; rather, it helps users understand how these technologies interact with imagery, what visual patterns they create, and how they influence modern culture.
Data Sources for Media
1. Social Media and Visual PlatformsTikTok, Pinterest, Instagram*, Behance, Are.na. — Examples of AI use in design, photography, fashion, and art. — Visual experiments and works by independent creators. — Discussions of emerging trends and audience reaction to AI-generated content.These sources are used to find topics and visual examples for publications.*Instagram is owned by Meta, a company designated as an extremist organization whose activities are prohibited in Russia
2. Professional Media and Thematic Platforms— It’s Nice That — publications covering contemporary design, illustration, and creative projects. — The Blueprint — materials on fashion, art, and visual culture. — Creative Applications — projects at the intersection of art and new technologies. — Dezeen — architecture, design, and technological experimentation.These are used to find interesting projects, interviews, original works, and examples of AI integration into creative industries.
3. Academic Research and Official Sources— arXiv — scholarly publications on generative models and computer vision. — OpenAI, Google DeepMind — research, documentation, and information regarding the development of generative technologies. — C2PA — materials concerning the provenance and authenticity of digital content.These sources are necessary for fact-checking, understanding how the technologies operate, and preparing analytical publications.
4. Editorial Original Content
The editorial team independently conducts experiments using generative tools, compares the results of various models, and investigates the unique characteristics of the generated imagery.
For instance, the editorial team asks several neural networks to create an image based on the exact same prompt, then compares the outcomes—examining composition, style, recurring motifs, and visual errors.
These experiments form the basis for the «HUMAN / AI» and «THROUGH THE EYES OF THE NEURAL NETWORK» columns.
5. Community MaterialsComments, user-submitted imagery, opinions from designers, photographers, and artists, and personal experiences interacting with generative tools.These community materials help us understand how professionals are perceiving changes in their fields, what challenges they are noticing, and what topics they wish to discuss.They form the foundation for discussions, interviews, and future editorial themes.
Data schema segmented for human and machine processing
Guidelines for Automated Services
AI receives the topic, column title, source materials, and specific assignment from the editor. The result of its work is a structured draft of the publication, complete with analysis, source citations, and notes on necessary verification.
Additional AI Usage Guidelines— Accuracy: Do not invent sources, authors, quotes, or facts. — Transparency: Indicate which parts of the material were created or processed using AI. — Authorship: Do not claim authorship over found images or projects. — Objectivity: Separate facts, observations, and interpretations. — Editorial Oversight: Do not make independent decisions regarding publication.
System example. Scenario: «Why do neural networks frequently generate identical images?»
01 Source Data — Editor
The Editor observes that social media frequently features AI-generated images sharing a similar aesthetic: glossy surfaces, unnatural lighting, idealized faces, and repetitive compositions. They compile a selection of such images from Pinterest, Behance, and Instagram* and present them to the AI with the task of identifying common visual techniques.
02 Machine Processing — AI
The AI analyzes the images and groups them based on visual characteristics: color palette, composition, lighting, textures, and style. It isolates recurring elements and generates a description of the similarities. Furthermore, it suggests possible reasons for the emergence of this aesthetic, distinguishing observations from hypotheses.
03 Results Review — Editor
The Editor verifies the identified patterns, studies the original images, and clarifies the context of their creation. They select the most illustrative examples to form a visual collection. In the caption, they explain which techniques are repeated and why they might create a feeling of uniformity. Unverified AI conclusions are excluded.
04 Outcome — Audience
The publication is released on Telegram in the «AI-SLOP DETECTOR» section as a carousel featuring image examples and corresponding commentary. Readers compare the works, notice the recurring techniques, and share their own examples. The experiment’s conclusion is a visual study that helps the audience realize the characteristics of generative image aesthetics and form their own opinion on the matter.
*Instagram belongs to Meta Platforms, Inc., a company designated as an extremist organization in Russia, and its activities are prohibited there.




