This paper explores the capabilities of publicly available artificial intelligence (AI) systems by analyzing their responses to a series of questions from the exact sciences and humanities. The questions relate to areas of science that, for various reasons, contain certain terminological, methodological, and other ambiguities, including those not always fully understood at a fundamental level. The questions touched, in particular, on concepts such as spontaneous and stimulated emission, the relationship between the meanings of the words “Great Russian” and “Russian”, and other related concepts. An analysis of the responses from the AI system named “Neuro” shows that its operation differs little from the process of selecting from the available elemental database a set of statements, definitions, etc., extracted from this database using the keywords contained in the original question. The AI is unable to detect logical violations in specific examples of terminology used in physics, or to identify the circumstances and causes of certain terminological perturbations in the humanities. The question of the completeness and consistency of the database it uses to formulate a final answer also remains unresolved. For these and other similar reasons, the AI under consideration cannot independently perform the functions of a normal scientific researcher. The question of the limits of AI improvement in terms of its ability to independently participate in scientific research as a fullfledged co-author is part of a larger fundamental problem that is still far from being resolved.
artificial intelligence; artificial intelligence; database; terms; logic; information warfare.