On the Creativity of AI Agents

2026-04-08 · Source: cs.AI updates on arXiv.org · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Robotics & Autonomous Systems · Depth: Expert, extended

Summary

This analysis explores the creativity of Large Language Models (LLMs), particularly when integrated into agentic systems, by proposing a dualistic framework. It differentiates between "functionalist creativity," which focuses on the observable characteristics of creative outputs like originality and effectiveness, and "ontological creativity," which emphasizes the underlying processes, social dimensions, and personal aspects of creation. The paper argues that LLM agents exhibit functionalist creativity, though not at its most sophisticated levels, but lack key aspects of ontological creativity, such as intrinsic motivation, continual learning, and intentionality. It also discusses the desirability of AI systems achieving both forms of creativity, evaluating potential benefits and risks, and suggesting pathways for artificial creativity to enhance human society, rather than merely substituting human roles.

Key takeaway

For AI Scientists and Research Scientists developing agentic systems, understanding the distinction between functionalist and ontological creativity is crucial. Focus your efforts on enhancing functionalist creativity by encouraging valuable divergence at learning and inference levels, and by developing new evaluation approaches that move beyond repetitive outputs. This approach can position AI as an augmentation tool, reducing ethical and legal risks while preserving human intellectual property and roles in creative domains.

Key insights

LLM agents demonstrate functionalist creativity but lack ontological creativity due to absence of intrinsic motivation and continual learning.

Principles

• Creativity requires both originality and effectiveness.
• Functionalist creativity focuses on observable output properties.
• Ontological creativity emphasizes process, social, and personal dimensions.

Method

The paper employs a dualistic framework to analyze AI creativity, separating it into functionalist (output-focused) and ontological (process-focused) perspectives to evaluate LLM agents.

In practice

• Develop new benchmarks for transformational creativity.
• Enhance reasoning and tool use with creative capabilities.
• Design effective ecosystems of interacting AI agents.

Topics

• LLM Agents
• Functionalist Creativity
• Ontological Creativity
• Transformational Creativity
• Intrinsic Motivation

Best for: AI Scientist, Research Scientist, AI Ethicist

Related on AIssential

• On the Creativity of AI Agents — LLM agents show functionalist creativity but lack ontological creativity, prompting debate on AI's creative potential.
• Under What Conditions Can a Machine Become Genuinely Creative? — Genuine machine creativity requires structural transformation of incomplete situations through recursive intervention, not just novel output.
• Scientists discover AI can make humans more creative — AI can enhance human creativity by acting as a collaborative partner, not just an automation tool.
• Automated Creativity Evaluation of Language Models Across Open-Ended Tasks — A new domain-agnostic framework automates large language model creativity evaluation across diverse open-ended tasks.
• "Imaginative AI" with Mohamed Elhoseiny — AI imagination can enable machines to perceive, create, act, and feel beyond their training data.
• CREATE: Testing LLMs for Associative Creativity — CREATE evaluates models' creative associative reasoning by generating diverse, specific concept paths.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by cs.AI updates on arXiv.org.