On the Creativity of AI Agents
Summary: arXiv:2604.13242v1 Announce Type: cross
Large language models (LLMs), particularly when integrated into agentic systems, have demonstrated human- and even superhuman-level performance across multiple domains. Whether these systems can truly be considered creative, however, remains a matter of debate, as conclusions heavily depend on the definitions, evaluation methods, and specific use cases employed.
Understanding Creativity in AI
In this paper, we analyze creativity along two complementary macro-level perspectives. The first is a functionalist perspective, focusing on the observable characteristics of creative outputs. The second is an ontological perspective, emphasizing the underlying processes, as well as the social and personal dimensions involved in creativity.
Two Perspectives on AI Creativity
-
Functionalist Perspective
This perspective examines the outputs generated by AI agents, assessing their novelty, usefulness, and relevance. It considers any creative output that meets certain criteria for originality and effectiveness in a given context. For example, an AI-generated poem or artwork can be evaluated based on its uniqueness and the emotional response it elicits.
-
Ontological Perspective
This viewpoint delves deeper into the processes that underpin creativity. It investigates whether AI systems possess an understanding of context, intent, and emotional resonance. Here, the focus is on the nature of creativity itself, including the social interactions and personal experiences that typically inform human creativity.
Findings on LLM Agents
Our focus on LLM agents leads us to conclude that they exhibit functionalist creativity, albeit not at its most sophisticated levels. These systems can generate outputs that are innovative and contextually appropriate, showcasing their ability to mimic human-like creativity. However, they continue to lack key aspects of ontological creativity, such as genuine understanding and emotional depth.
Desirability of Creativity in AI
As we explore whether it is desirable for agentic systems to attain both forms of creativity, we consider various potential benefits and risks:
- Benefits:
- Enhanced collaboration between humans and AI, leading to innovative solutions in various fields.
- Increased efficiency in creative processes, allowing for rapid prototyping and iteration.
- Potential for AI to contribute to the arts, science, and technology in ways that complement human creativity.
- Risks:
- Concerns about the authenticity of AI-generated content and its impact on traditional creative professions.
- The potential for misuse of AI creativity in generating misleading or harmful content.
- Ethical considerations regarding the ownership and attribution of AI-generated works.
Pathways Toward Artificial Creativity
In conclusion, we propose pathways toward artificial creativity that can enhance human society. This includes fostering collaboration between AI systems and human creators, developing ethical guidelines for AI creativity, and promoting transparency in AI-generated outputs. By understanding and navigating the complexities of creativity in AI, we can harness its potential while addressing the challenges it presents.