AI vs Human-Generated Historical Game Narratives Across Three Cultural Settings

A 2026 study compares narrative performance between 10 advanced AI models and 30 professional game designers. The research employs a two-stage empirical design, generating and evaluating narratives set in Ancient Egypt, Medieval Europe, and Ming Dynasty China. Expert evaluation used a validated five-dimensional framework covering historical authenticity, interactivity, immersion, structure, and innovation.

Use Cases

• Benchmarking AI narrative generation quality based on a five-dimensional evaluation framework
• Studying the impact of cultural familiarity on narrative quality based on the three historical settings
• Designing hybrid AI-human creative workflows based on the identified complementary strengths

Strengths

• Empirical design involving 10 AI models and 30 human designers
• Evaluation based on a validated five-dimensional framework (historical authenticity, interactivity, immersion, structure, innovation)
• Statistical analysis indicates approximately 70% of evaluated AI models achieved or exceeded average human performance in overall scores

Limitations

• Row count is unknown, which may limit suitability assessment
• Column-level documentation is absent; field semantics must be inferred after download
• Data may reflect bias inherent to the specific AI models and designer pool used

Provenance

Source

Alyssa Xiaoming Dai via figshare

Collection Method

Two-stage empirical study involving narrative generation and expert evaluation

Freshness

Last updated 2026-05-13 08:58:37; freshness should be verified

Geography

Narratives set in Ancient Egypt, Medieval Europe, and Ming Dynasty China