Cambridge researchers found that simply adding AI to creative tasks does not automatically improve results. The study, published in Information Systems Research, involved three linked experiments, each with 160 to 200 participants.
Human-AI pairs did not naturally become more creative through repeated collaboration alone, and some became less creative over time. The researchers examined three collaboration approaches i.e.: humans proposing ideas, humans asking AI to generate ideas, and humans and AI jointly refining ideas.
Creativity improved only when participants focused on co-developing ideas, exchanging feedback and building on suggestions, rather than continuously generating new ideas without refinement.
A third experiment showed that explicit instruction to engage in co-development led to clear creativity gains across repeated tasks.
Dr. Yeun Joon Kim of Cambridge Judge Business School stated that organizations must provide targeted support, such as guidance on building and adapting ideas, to help employees and AI learn over time how to create more effectively.
The research indicates that companies should structure AI collaboration through clear instructions and workflows, rather than relying on AI use alone, to improve creative outcomes.
Notes: This post was drafted with the assistance of AI tools and reviewed, edited, and published by humans. Image: DIW-Aigen
Read next:
• How Much Do Fake Account Verifications Really Cost Across the World?
• Google Translate Gets AI Upgrade: Gemini Now Powers Text and Speech Translation
Human-AI pairs did not naturally become more creative through repeated collaboration alone, and some became less creative over time. The researchers examined three collaboration approaches i.e.: humans proposing ideas, humans asking AI to generate ideas, and humans and AI jointly refining ideas.
Creativity improved only when participants focused on co-developing ideas, exchanging feedback and building on suggestions, rather than continuously generating new ideas without refinement.
A third experiment showed that explicit instruction to engage in co-development led to clear creativity gains across repeated tasks.
Dr. Yeun Joon Kim of Cambridge Judge Business School stated that organizations must provide targeted support, such as guidance on building and adapting ideas, to help employees and AI learn over time how to create more effectively.
The research indicates that companies should structure AI collaboration through clear instructions and workflows, rather than relying on AI use alone, to improve creative outcomes.
Notes: This post was drafted with the assistance of AI tools and reviewed, edited, and published by humans. Image: DIW-Aigen
Read next:
• How Much Do Fake Account Verifications Really Cost Across the World?
• Google Translate Gets AI Upgrade: Gemini Now Powers Text and Speech Translation
