Researchers Automate Creativity Assessment with AI

A team from South Ural State University and the Russian Academy of Sciences' Institute of Psychology has developed an artificial intelligence system designed to evaluate human creativity. Their work centers on automating the scoring of established projective tests, such as the Torrance Tests of Creative Thinking.

The project began by compiling Torrance test results from three thousand Russian participants. Researchers used this dataset to train a neural network model. The system now processes diagnostic results from multiple creativity assessments, delivering an analysis in minutes—a task that typically requires significant time from a trained psychologist.

A dedicated online portal already allows individuals to take creativity tests based on both the Torrance and Urban methodologies. Users receive scored results almost immediately, without needing a consultation. The university's model expands on the six standard Torrance metrics—like originality, fluency, and resistance to premature closure—by incorporating supplementary indicators for cultural context and personal attributes.

This development follows other recent applications of neural networks for analytical tasks in Russia. For instance, emergency services are testing a similar system to accelerate fire investigations. That model analyzes images from a blaze, identifies key objects, and produces a formalized report in seconds, replacing a manual process that often takes an hour.

Source: Lenta.RU