Limitations on the use of Large Language Models imposed by information age

The paper investigates the limitations on application of Large Language Model due to information age. Information age is understood as the period between successive data updates. If the data has been updated later than the time of model training, requesting such information from the model will not be valid. In most cases, such queries are redirected to the search engine. Therefore, the structure of a response to queries related to recent events differs from the standard output. This feature can be used in the Turing test for recognizing artificial intelligence. The paper presents experiments to determine the data lifetime and moment of model training.