AI Research Support Systems - Taxonomy

This taxonomy emerges from the need to systematically map and analyze AI-based Research Support Systems, with a specific focus on tools designed to assist key phases of the research lifecycle. In particular, it targets systems that support literature discovery, concept mapping, gap detection, and automatic citation generation, which are increasingly being enhanced through the integration of LLMs. The rapid proliferation of such tools calls for a structured framework to better understand their functional capabilities, architectural choices, and underlying data strategies. Beyond functionality, this taxonomy also investigates the context of development, distinguishing between academic, industrial, and hybrid initiatives, in order to highlight differences in design goals, openness, and reproducibility. A central aspect of this analysis concerns the adoption of Knowledge Graphs (KGs). These structured data representations play a crucial role in enabling improved transparency, verifiable reasoning mechanisms, and mitigation of LLM hallucinations through grounded information. The taxonomy explicitly captures whether and how systems leverage KGs, distinguishing between open graph, proprietary, and dynamically generated approaches and if they are domain-specific or general-purpose.