Short answer

Embrace the 'uncanny' in AI design: aim for outputs that are intelligently surprising and offer a unique, non-human perspective, rather than just mimicking human creativity.

Field: Innovation & Design
Source: Journal of Artificial General Intelligence (2015)
Method: Qualitative analysis and case study
Evidence: Moderate effect

AI systems can be perceived as more creative when they exhibit a blend of human-like understanding and non-human detachment, rather than attempting to perfectly mimic human output. This innovation & design research insight is drawn from a 2015 study published in Journal of Artificial General Intelligence. Using Qualitative analysis and case study, researchers explored how this design variable affects real-world outcomes. The key design takeaway: Embrace the 'uncanny' in AI design: aim for outputs that are intelligently surprising and offer a unique, non-human perspective, rather than just mimicking human creativity.

Study

Innovation & DesignHigh ImpactModerate effect

The Uncanny Valley of AI: Balancing Human-like Output with Non-Human Novelty

AI systems can be perceived as more creative when they exhibit a blend of human-like understanding and non-human detachment, rather than attempting to perfectly mimic human output.

Journal of Artificial General Intelligence · 2015

Key Findings

01Human expectations of AI outputs increase with the sophistication of the systems.
02AI systems do not need to perfectly replicate human creativity to be perceived as creative; a degree of 'unnaturalness' can enhance this perception.
03The 'uncanny' – the feeling evoked by objects at the boundary of human and non-human categories – can be a source of perceived creativity in AI.
04Twitterbots that combine meaningful output with a non-human perspective can be popular and engaging.

Application

Design takeaway

Embrace the 'uncanny' in AI design: aim for outputs that are intelligently surprising and offer a unique, non-human perspective, rather than just mimicking human creativity.

How to apply

When designing AI-powered creative tools, consider incorporating elements that deliberately introduce unexpected juxtapositions or insights that a human might not readily conceive, stemming from the AI's unique processing capabilities.

Project actions

01Explore how your AI system's unique computational processes can lead to outputs that are novel and unexpected.
02Consider how to present AI-generated content in a way that highlights its non-human origin while still being relatable.

Method & Evidence

AimHow can generative AI systems be designed to elicit perceptions of creativity by balancing human-like qualities with distinct non-human characteristics?

MethodQualitative analysis and case study

ProcedureThe research examines the concept of the 'uncanny' in relation to AI-generated content, using a Twitterbot (@MetaphorMagnet) as a case study to explore how systems that blend human understanding with non-human perspectives can be perceived as creative.

ContextGenerative AI, Human-Computer Interaction, Digital Art and Content Creation

Variables

IVDegree of human-like mimicry vs. non-human novelty in AI output.

DVPerceived creativity of AI output.

CVDomain of AI generation (e.g., text, image), user familiarity with AI.

Strengths & Limitations

Strengths

+Introduces a nuanced perspective on AI creativity beyond simple imitation.
+Provides a theoretical framework (the uncanny) for understanding user reactions to AI.

Limitations

The subjective nature of 'creativity' and 'uncanny' makes objective measurement challenging. The specific domain of text generation might not apply universally.

Reliability & validity

Reliability could be improved by using a larger and more diverse sample of participants. Validity is supported by the theoretical grounding in psychological concepts like the uncanny, but subjective ratings of creativity can be a limitation.

Think critically

To what extent should AI be designed to 'hide' its non-human nature, versus embracing it as a source of unique creative output?

Design Principles

"The creative potential of artificial intelligence is amplified by its ability to synthesize human-like understanding with distinct, non-human perspectives, creating a novel 'uncanny' effect."

Understanding the 'uncanny' aspect of AI-generated content is crucial for designers developing AI tools. It suggests that striving for perfect human replication might be less effective than creating systems that offer novel perspectives or surprising juxtapositions, leveraging their non-human nature.

What This Means for Your Design

AI can be seen as more creative if it's a bit weird and surprising, not just if it acts exactly like a person. Think of it like a robot that can write poetry – it's cool because it's a robot doing something human, but in its own unique way.

How to use in your project

1.Reference this study when discussing how user expectations of AI influence their perception of its outputs, particularly in the context of creativity and novelty.

Add to My Project

Quick Cite

Paragraph starter

The perception of creativity in generative AI is influenced by a balance between human-like understanding and distinct non-human characteristics. Research suggests that AI systems can be perceived as more creative when they exhibit an 'uncanny' quality, blending familiar human concepts with novel, detached perspectives, rather than attempting perfect human mimicry (Veale, 2015). This implies that designers should explore how to leverage AI's unique computational nature to produce surprising and thought-provoking outputs.

Source

Journal of Artificial General Intelligence

Unnatural Selection: Seeing Human Intelligence in Artificial Creations

journal · 2015

View source

Questions About This Research

What does the research say about the uncanny valley of ai: balancing human-like output with non-human novelty?: Embrace the 'uncanny' in AI design: aim for outputs that are intelligently surprising and offer a unique, non-human perspective, rather than just mimicking human creativity. Evidence: Journal of Artificial General Intelligence (2015).
Why does "The Uncanny Valley of AI: Balancing Human-like Output with Non-Human Novelty" matter for design?: Understanding the 'uncanny' aspect of AI-generated content is crucial for designers developing AI tools. It suggests that striving for perfect human replication might be less effective than creating systems that offer novel perspectives or surprising juxtapositions, leveraging their non-human nature.
How can designers apply this research?: Embrace the 'uncanny' in AI design: aim for outputs that are intelligently surprising and offer a unique, non-human perspective, rather than just mimicking human creativity.
What were the main findings?: Human expectations of AI outputs increase with the sophistication of the systems.. AI systems do not need to perfectly replicate human creativity to be perceived as creative; a degree of 'unnaturalness' can enhance this perception.. The 'uncanny' – the feeling evoked by objects at the boundary of human and non-human categories – can be a source of perceived creativity in AI.. Twitterbots that combine meaningful output with a non-human perspective can be popular and engaging.
What research method was used?: Qualitative analysis and case study.
How strong is the evidence?: Evidence strength is rated Moderate effect, based on a 2015 journal from Journal of Artificial General Intelligence.
What should I do differently in my next project?: When designing AI-powered creative tools, consider incorporating elements that deliberately introduce unexpected juxtapositions or insights that a human might not readily conceive, stemming from the AI's unique processing capabilities.
What are the limitations?: The study focuses on text-based AI (Twitterbots) and may not directly translate to other AI modalities. The perception of 'uncanny' and 'creativity' is subjective.