Algorithm-driven tools hold the promise of democratizing knowledge access and boosting creativity. However, research shows a hidden downside: these tools can restrict creative potential by undervaluing expertise. The root cause lies in their design. Algorithms typically prioritize popular and relevant information, reinforcing users’ existing knowledge rather than encouraging new exploration. This creates ‘ideation bubbles’—clusters of similar ideas leading to homogeneity in thinking.
Our research introduced an alternative approach, modifying traditional algorithms to emphasize diversity and uncommon ideas. In experiments involving sustainability challenges, users of this exploration-focused algorithm produced notably more creative solutions. Experts, in particular, benefited, outperforming novices by leveraging diverse insights and combining knowledge across fields—a process we call recombinant innovation.
The organizational impact is significant: exploration-based algorithms help experts break free from conventional thinking, generating a broader range of novel ideas. Businesses should treat algorithm design as a strategic choice, matching exploration tools to innovation tasks while maintaining exploitation tools for efficiency. Encouraging experts to critically engage with algorithm-generated content and continuously auditing idea diversity can further combat ideation bubbles. Ultimately, expertise remains essential, but its value transforms with AI, shifting toward the ability to synthesize and innovate from diverse information pools.
This shift calls for investing in expertise and configuring AI tools to unlock creative potential, positioning organizations to harness AI-driven innovation effectively.
