We introduce NeoN, a new system designed to detect and conduct preliminary analyses of newly emerging words in Polish. Developed by the Linguistic Engineering Group at the Institute of Computer Science of the Polish Academy of Sciences, NeoN combines corpus and dictionary checks, frequency analysis, contextual lemmatization, and orthographic normalization with a Large Language Model module. This multi-layer architecture effectively processes daily updated text sources (e.g., RSS feeds) to accurately identify lexical innovations while filtering out problematic/error cases.

During our presentation, we will outline the rationale and design principles behind NeoN, demonstrate the interface and its context display features, and illustrate its functionalities, such as automated categorisation and definition generation. We will also present concrete examples of newly detected lexical items and highlight how a human-in-the-loop approach is necessary for making good use of automated outputs. Finally, we will discuss potential future developments, such as extending NeoN to other languages and integrating additional functionalities to further support the automated monitoring and analysis of lexical innovation.