DeepSlide AI System Enhances Presentation Delivery with Human-AI Collaboration

A new AI system called DeepSlide, described in a May 2026 arXiv preprint titled ‘DeepSlide: From Artifacts to Presentation Delivery,’ optimizes both creation and delivery of academic presentations through human-AI collaboration. The system addresses gaps in current AI-generated slide tools by incorporating presentation dynamics alongside visual design.

The preprint details DeepSlide’s time-budgeted narrative planning, evidence-grounded content retrieval, and rehearsal support features. This architecture enables iterative refinement through human feedback during the planning process to improve pacing and narrative flow.

Traditional AI slide generators focus primarily on visual design, but DeepSlide expands optimization to include delivery mechanics. The paper notes this dual focus could enhance scholarly communication outcomes by improving both artifact quality and presentation effectiveness.

The research is available through arXiv, a US-based academic repository. The work advances AI-assisted presentation tools for academic and professional settings requiring structured knowledge delivery.