Art, algorithms and anatomy: How researchers can partner with AI without losing the human touch

Generative artificial intelligence is shaking up the academic world, sparking fears that automation will make researchers lazy and strip out the deep, contextual meaning behind qualitative analysis. But a new study suggests that instead of replacing human intellect, AI can be treated like an artistic collaborator to spark unexpected discoveries.

The study, published in Strategic Organization, draws on an unusual blend of fine art history and modern data analysis to show how large language models (LLMs) can help scholars find fresh insights.

"Thoughtfully integrated AI can enhance qualitative research by promoting discovery and surprise, both essential elements of theory building," says Vern Glaser, professor at the Alberta School of Business.

To build their framework, Glaser and co-author Jennifer Sloan (‘25 PhD) at University College of London, looked to the creative processes of two famous artists: Pablo Picasso and digital media artist Refik Anadol.

The researchers used Picasso’s 1945 lithograph series The Bull — where the artist progressively stripped away physical details until the animal was reduced to a few essential, elegant lines — as a metaphor for data reduction. Conversely, they looked at Anadol’s Unsupervised exhibition at the Museum of Modern Art, which synthesizes millions of data points into flowing digital art, as a model for data synthesis.

By treating data reduction and data synthesis as "complementary engines of insight," the study identifies four distinct pathways where AI can surprise researchers into seeing their data in entirely new ways. These include multiplying theoretical lenses, surfacing unmentioned absences in the data, bridging micro and macro levels of analysis, and testing categories against hybrid data points.

However, the authors warn that using AI blindly is dangerous. Without strict human oversight, the technology can smooth over contextual nuances, flatten data or generate "colourful noise" that sounds smart but holds no true scientific weight.

To prevent this, the study outlines a concept they call "interpretive vigilance". This framework requires researchers to remain the absolute gatekeepers of meaning, treating AI outputs purely as creative proposals rather than proven facts.

"Meaning-making remains squarely in human hands. Our framework positions AI as a collaborative partner that amplifies researchers' capacity for theoretical discovery while preserving methodological rigor and interpretive depth."

— Vern Glaser

Key takeaways

• AI as a creative partner: Instead of automating away human thought, generative AI can be used in academic research to find unexpected patterns and juxtapositions that humans might overlook.
  
  
• The art of data: Effective data analysis requires a balance of reduction (simplifying complex data, like Picasso's drawings) and synthesis (combining massive datasets into new forms, like digital media art).
  
  
• Four surprise pathways: Researchers can use AI to look through multiple theoretical viewpoints at once, spot missing information or silences in text, connect small details to bigger systemic trends, and test out boundaries of their data categories.
  
  
• Interpretive vigilance: Human scholars must remain strictly critical of AI findings, verifying every automated suggestion against real field evidence and maintaining a clear, transparent history of their AI prompts.

Read the full article in the journal Strategic Organization at DOI:10.1177/14761270261448648

U of A research. Real-World impact.

Bold ideas in action! From reimagining industries to building sustainable infrastructure, the U of A is helping to achieve the UN's goal of creating a more resilient, tech-powered world.

See How We're Igniting Innovation