Lexical Ambiguity in AI-Generated Texts: Linguistic Challenges and Limitations

This article examines lexical ambiguity in texts produced by generative artificial intelligence. The analysis focuses on polysemy, homonymy, insufficient contextual clues, and semantic shifts between sentences. The study employs a qualitative approach based on original examples specifically constructed for linguistic analysis. These examples demonstrate how grammatically correct sentences may allow multiple interpretations. Particular attention is given to words frequently used in academic and professional communication. The article also compares the interpretation of selected English words in Russian and Azerbaijani. This comparison demonstrates that accurate translation depends on context rather than a single dictionary equivalent. To reduce ambiguity, writers are encouraged to clarify key terms and maintain terminological consistency throughout their texts. Teachers can help students develop semantic awareness and critical reading skills through comparison and editing activities. The study concludes that AI-generated texts require human review and that the proposed approach should be further evaluated using authentic AI-generated materials.