Towards Defining, Measuring, and Leveraging Creativity in AI Applications

URL SCAN: Towards Defining, Measuring, and Leveraging Creativity in AI Applications
FIRST LINE: # Computer Science > Artificial Intelligence

The Dissection

The paper translates creativity from a supposedly irreducible human quality into a score built from newness, value, and surprise. Its validation, as described, tests whether the formula captures recognizable patterns in chess moves and paintings. That is measurement of judged artifacts—not proof of creative agency, generative originality, scientific discovery, or lasting cultural value.

The Core Fallacy

It treats criteria as mechanisms. Newness, value, and surprise are not independent physical variables; they are judgments dependent on reference sets, evaluators, purpose, and historical context. A formula that reproduces human assessments is formalized evaluation, not creativity itself. Domain-specific adaptation may make the metric useful, but it weakens the claim of a genuinely unified structure.

Hidden Assumptions

• Creativity can be reduced to a single scalar without destroying its meaning.
• Value can be measured before downstream adoption, use, or historical validation.
• Surprise can be separated from ignorance, complexity, or statistical rarity.
• Metrics developed for chess and painting generalize to science and other creative domains.
• Human labels and cultural priors are not silently embedded in the measurements.
• Capturing patterns in existing works is equivalent to measuring creativity.

Social Function

Classification: partial truth, prestige signaling, and transition management.

The paper contains a real insight: creative judgments often recur around novelty, value, and surprise. But its deeper social function is to make a human status claim machine-readable and therefore optimizable. It turns creativity into an evaluation pipeline for AI systems. If marketed as proof that creativity remains uniquely human, it becomes ideological anesthetic. The mechanism runs in the opposite direction: formalization strips away the mystery that could function as a labor moat.

The Verdict

This paper does not refute the Discontinuity Thesis. It supplies infrastructure for it. If the formula is useful, it advances P1 by making creative output searchable, rankable, and optimizable. The abstract does not establish P1–P3 by itself; it provides no evidence of AI cost superiority, institutional coordination failure, or mass labor displacement. But its trajectory is unambiguous: creativity is being decomposed into measurable components, not rescued as a human sanctuary. Useful evaluation scaffold. Weak universal theory. Potentially valuable transition technology for automating creative work.