Highlights:
- Researchers demonstrate that personalization significantly enhances human-AI synergy in creative tasks.
- Personalized AI systems leverage psychometric profiles and user style interviews to co-create marketing content.
- Study involving 331 participants reveals higher creativity, trust, and collaboration with personalized AI models.
- Causal mediation analysis links personalization to improved memory, reasoning, and joint cognition.
TLDR:
A new study by Sean Kelley, David De Cremer, and Christoph Riedl introduces a personalized AI framework that adapts to users’ psychological and professional profiles, enabling more effective and creative human-AI collaboration. The findings suggest personalized large language models could redefine AI’s role in supporting creative work.
In a major breakthrough for the field of Human-Computer Interaction, researchers Sean Kelley, David De Cremer, and Christoph Riedl have unveiled a novel AI framework designed to enhance creative collaboration between humans and machines. Their paper, titled ‘Personalized AI Scaffolds Synergistic Multi-Turn Collaboration in Creative Work,’ explores how tailoring AI systems to individual cognitive and behavioral traits can substantially improve both productivity and creativity in knowledge work. As artificial intelligence becomes more entwined with daily workflows, the ability for machines to understand and adapt to human collaborators is emerging as a key challenge for the next generation of intelligent systems.
The team conducted a large-scale experiment with 331 participants, dividing them into three groups that interacted with different levels of AI personalization. The first group worked with a generic large language model (LLM), the second with a partially personalized version, and the third with a fully personalized assistant informed by users’ psychometric assessments and self-reported work styles. The results were decisive: participants who collaborated with fully personalized AI produced marketing campaigns that were significantly higher in creativity, coherence, and originality. Participants also reported greater trust in the AI, along with higher levels of satisfaction, perceived feedback quality, and a sense of collaborative partnership. These outcomes point toward a paradigm shift in human-AI interaction, emphasizing the value of mutual understanding and adaptive co-creation.
Technically, the researchers’ framework uses a combination of psychometric profiling, user interviewing, and iterative fine-tuning of the LLM’s response strategies. This approach enhances what the authors call ‘collective memory’—the AI’s ability to retain and build upon prior interaction context—alongside improved attention and reasoning in multi-turn dialogues. By functioning as an external cognitive scaffold, personalized AI reduces uncertainty in communication and fosters a shared understanding of goals and tasks. The study’s causal mediation analysis supports this model, demonstrating that personalized assistance indirectly improves creative performance by strengthening the alignment between human intention and AI reasoning processes. The implications of this research extend far beyond marketing, suggesting broad applications for education, design, consulting, and scientific innovation.
Ultimately, Kelley, De Cremer, and Riedl’s work presents a theoretical and practical foundation for developing AI assistants that amplify human creativity while avoiding the risks of homogenized or one-size-fits-all models. As organizations look to leverage AI for competitive advantage, this approach highlights that personalization—rooted in empathy, cognition, and mutual adaptation—may be the key to achieving true human-AI synergy.
Source:
Source:
arXiv:2510.27681v1 [cs.HC] (https://arxiv.org/abs/2510.27681) by Sean Kelley, David De Cremer, and Christoph Riedl, submitted on October 31, 2025.