Human-AI Taste Collaboration: Interactive Scientific Idea Co-Creation

TL;DR: Let’s pair humans and AI to brainstorm research ideas, seeing if their combined “taste” produces better science than either alone. We’ll measure if AI can enhance—not just mimic—human creativity in science.

Research Question: How does interactive, real-time collaboration between humans and AI models trained for scientific taste influence the quality and originality of research idea generation?

Hypothesis: Collaborative systems that blend human intuition and AI-derived taste judgments can outperform both solo AI and solo human brainstorming, especially in generating ideas that balance novelty and feasibility.

Experiment Plan: Develop an interactive research idea co-creation platform where users can generate, critique, and refine ideas with an AI partner (powered by Scientific Thinker). Run controlled experiments with groups of researchers, measuring the quality, originality, and downstream impact of collaboratively generated ideas versus those from humans or AI alone. Analyze logs to identify patterns of synergy or divergence in taste between human and machine.

References:

• Tong, J., et al. (2026). AI Can Learn Scientific Taste.
• Mogili, V. B. (2025). Empowering Human-AI Collaboration: Enterprise Technology Platforms and Human Expertise Synergy in Healthcare, Finance, and Scientific Research. Journal of Computer Science and Technology Studies.
• Volcu, I. (2025). Creative attitudes - important aspect in the training process of specialists in the field of physical education and sport. "Sports. Olimpysm. Health", international scientific congress.