We investigate whether immediate repetition of nonsense words can distinguish cognitively healthy adults from those with mild cognitive impairment (MCI) or dementia. In a computer-based study, 129 Czech speakers (45–84 y) repeated six pseudowords; each session was recorded and transcribed by four state-of-the-art ASR models. Grapheme-level similarity between the transcript and the target word served as a phonological accuracy score. Using logistic regression, the best ASR variant (wav2vec-nolm, no language model) separated patients from controls with 77 % accuracy, whereas language-model–’corrected’ hypotheses performed markedly worse. These findings show that a very short, vocabulary-free task can reveal early linguistic decline while requiring only consumer hardware and minimal instruction. Incorporating such a nonsense word repetition subtask into digital neuropsychological batteries could therefore sharpen large-scale, remote screening and help clinicians focus full assessments on individuals at greatest risk.