Assessing Emotion and Sensitivity of AI Artwork

Area of Science:

• Cognitive Science
• Artificial Intelligence Ethics
• Art Perception

Background:

• Artificial Intelligence (AI) is increasingly creating art, a domain previously exclusive to humans.
• Empirical data on public perception of AI's creative capabilities is limited.
• Understanding human biases towards AI-generated content is crucial.

Purpose of the Study:

• To investigate how knowledge of an artwork's AI origin affects human perception and evaluation.
• To determine if stereotypes about AI creativity influence aesthetic judgments.
• To provide empirical evidence on the impact of AI in creative domains.

Main Methods:

• Two experiments exposed participants to AI-generated art (audiovisual and auditory).
• Participants evaluated artwork, with conditions differing in whether the artist was disclosed as human or AI.
• Data collection included evaluations of sensitivity, emotional impact, and overall quality.

Main Results:

• Participants rated AI-generated art lower in sensitivity, emotional evocation, and quality when aware of its AI origin compared to when believing it was human-created.
• Results were consistent across different art forms and experimental settings.
• Art evaluation is significantly modulated by pre-existing stereotypes and biases regarding AI's creative skills.

Conclusions:

• Human perception of art is influenced by preconceived notions about the creator's nature (human vs. AI).
• Biases against AI's creative potential can lead to diminished aesthetic appreciation.
• Further research is needed to understand and mitigate these biases in human-AI creative collaborations.