Shaping changes, resonating research
Artistic research today encounters the critical challenge of extending beyond its immediate context and bridging disciplinary divides. In a landscape where art, culture, and both local and global policy intersect within complex societal dynamics, the question arises: how can artistic researchers actively contribute to shaping these dialogues? We propose that addressing this question first requires understanding how artistic research can truly resonate beyond its specific contexts and disciplinary frameworks.
Bridging semantic content
°'°Kobi functions to dissolve disciplinary silos by leveraging AI to process and integrate a diverse corpus of artistic and scientific publications into a connected knowledge ecosystem. Within this system, each knowledge fragment—whether a text, image, or media element—is assigned a semantic position in an interdisciplinary knowledge universe. Through advanced analytical processes, the AI identifies core concepts, keywords, and titles, establishing meaningful links between items based on thematic resonance.
How KOBI Works:
1. Importing the semantic context: A semantic context is imported, which in our case is represented by an exposition from the Research Catalogue.
Future Developments
● Application of machine learning algorithms to the knowledge base to optimize learning processes.
The °'°Kobi System
To foster this resonance, we introduce °'°Kobi, an AI-driven research platform designed to build connections across diverse fields and stimulate expansive, creative inquiry. Here, resonance is understood as the creation of a dynamic relationship between artistic research and the broader world—one that supports reflection, exchange, and a capacity for mutual transformation. As art and culture engage with increasingly interconnected challenges, there is a growing need for tools that transcend disciplinary boundaries to foster collaborative exploration. °'°Kobi is conceived to meet this need, establishing spaces for interdisciplinary dialogue and intellectual cross-pollination.
Creative Divergence
For users, an AI chatbot manages interactions, responding to inquiries with contextually rich insights drawn from the knowledge base. These responses encourage users to explore alternative perspectives, uncover unexpected connections, and engage deeply with ideas across disciplines, thus fostering a fertile ground for interdisciplinary inquiry and creative engagement.
Future Developments
● Creation of a specialized glossary to facilitate the understanding and use of key terms.
Future Developments
● Expansion and enhancement of the knowledge base to ensure more comprehensive and accurate content for users.
The °'°Kobi System
°'°Kobi is a Knowledge Management System conceived as a vast associative memory. Currently, its knowledge base is integrated with the Research Catalogue, a database dedicated to artistic research structured into expositions. In °'°Kobi, an exposition represents a complex unit of knowledge that combines text (with a mandatory abstract), images, and sounds, offering an articulated presentation of the results of artistic research.
Artificial Intelligence has opened up new scenarios and opportunities in the field of learning, transforming traditional approaches and prompting a reflection on how to renew educational methods. In the context of the emerging debate in our country on enhancing infrastructures and tools dedicated to artistic research, the Academy of Fine Arts in Rome, in collaboration with the Marche Polytechnic University, developed °'°Kobi, an AI-mediated platform aimed at fostering critical and divergent thinking.
In a historical moment characterized by increasingly unpredictable progress, the British educator Ken Robinson argued that creativity would become as important and necessary as literacy, a thought encapsulated in his famous 2006 TED talk Do schools kill creativity? That same year, AI@50 was held in Dartmouth, a conference organized to celebrate the “golden anniversary” of artificial intelligence, marking a significant step in the historicization of AI by highlighting its advancements and the maturity of the technology, which was now ready to initiate collaborations with other disciplines. The need to encourage creativity in educational paths, alongside the parallel development of intelligent systems, laid the foundation for studies and research that leverage the adaptability of new technologies for educational purposes.
The system presents itself not only as a consultation tool but as a means of making knowledge usable, facilitating its circulation and connection, and promoting creative research.
How KOBI Works:
7 Knowledge Connection:
by iterating the previous steps, the system generates connections that were not present in the original material. These connections link concepts or keywords found in different contexts.
Future Developments
● System proactivity, offering research suggestions based on the divergence of queries.
Even before any interaction with the user, the machine autonomously performs a connection function, reconfiguring the databases or knowledge archives.
How KOBI Works:
4. Paragraph analysis and titling: The text is divided into paragraphs, each analyzed individually by the AI, which also generates a title for each.
How KOBI Works:
3. Lemmatization of concepts: Key concepts are lemmatized, generating a set of keywords to facilitate categorization and search.
°'°Kobi presents itself to the user with two different interfaces: a chatbot, which allows for a conversation on a topic proposed by the user, and “Universe,” a constellation of thematic and conceptual nodes designed to be explored even in augmented reality. The Universe is characterized by spheres of different sizes and colors, which respectively distinguish semantic contexts, concepts, and keywords, interconnected by a network of lines created by the principle of semantic proximity. With these two exploratory modes, °'°Kobi offers an experience that informs and deconstructs, leading to unexpected places and expanding the horizon of discovery and learning.
The AI-mediated educational environment promotes a non-linear learning process, acting as a catalyst for divergent thinking, one of the main indicators used to identify and measure creativity. Unlike convergent thinking, which focuses on finding a single correct solution or the best choice among predefined options, divergent thinking encourages an approach that explores unusual alternatives, new and original ideas—concepts that form the basis of the standard definition of creativity.
The etymological choice, derived from the combination of the Greek terms “koinos” (κοινός), meaning “common” or “shared,” and “bios” (βίος), meaning “life” or “existence,” reflects an educational approach that embraces the notion of an ecosystem applied to learning contexts and digital archiving based on collective intelligence. Indeed, students, under the supervision of professors, can actively contribute to expanding °'°Kobi by creating projects within the Research Catalogue, a database that forms the platform’s knowledge base through a collaboration with the Society for Artistic Research.
How KOBI Works:
8. Navigation: The exploration of the semantic space makes every action within the knowledge space measurable; specifically, the divergence of different interactions can be measured.
The dual nature of artistic education, which combines the study of theoretical disciplines like art theory and aesthetics with the development of skills and competencies through experience and practice, is now influenced by artificial intelligence technology. Beyond tools that utilize so-called 'artificial creativity' or 'pseudo-creativity,' concepts defining the unique creative abilities of machines, certain AI models are proving particularly useful in enhancing associative and combinatory creativity, essential in the early exploratory phases of an artistic project. This is the case with LLMs (Large Language Models), advanced AI models capable of understanding and generating natural language, which are integrated into the technological infrastructure of °'°Kobi, whose name explicitly reflects the platform’s nature and purpose.
As an ecosystem fueled by the same artistic community that uses it, °'°Kobi proposes a paradigm of collective intelligence and creativity where processes are materially and temporally distributed, accommodating the needs and modes of a society interconnected and immersed in online realities. For digital natives, as highlighted by the entry “Creativity in the Future” from the Encyclopedia of Creativity (2020), creativity is no longer identified with the romantic image of the solitary and extraordinary genius but is rather a daily, collective, and relational activity, accessible to all and experienced everywhere.
How KOBI Works:
2. Extracting key concepts: Through one of the most interesting emerging features of LLMs, key concepts are extracted from the exposition’s abstract.
Once the concepts and keywords have been assigned a unique position within the semantic space, all other elements of the original semantic context are associated based on proximity. Each element thus becomes a knowledge node, contributing to the construction of an interconnected network.
EAR: Enacting Artistic Research – Research as a key factor for the internationalization of AFAM institutions
(MUR PNRR 2023, ID INTAFAM00060, funded by Next Generation EU)
In addition to being a system of interactions and sedimentation of memories, digital and biological, individual and collective, °'°Kobi allows for semantic exploration that facilitates a non-conventional cognitive process, open to multiple directions. By using the principle of proximity, the system identifies similarities and polar opposites by grouping related information, facilitating the organization, interconnection, and interpretation of complex ideas and concepts.
How KOBI Works:
6. Creating the new context and semantic space: The content of the exposition, already analyzed and transformed into a constellation of knowledge atoms, is placed in a semantic space. For this step, one of the pre-existing models is used, which assigns a unique position to keywords and concepts. Kobi uses Chroma DB, a vector space with about 400 dimensions.
How KOBI Works:
5. Evaluation of media materials: Media materials are examined individually, but only in relation to their semantic proximity to the text. Their specific content is neither analyzed nor translated through descriptors.
Responding to Francesco d’Isa’s invitation to see artificial intelligence as a mirror, we present below a summary that outlines the internal structure of °'°Kobi, allowing the reader to undertake a meta-reading that captures the overlaps and dissonances between the processes that animate artistic creation and those that underlie the °'°Kobi knowledge ecosystem.