If you are using AI-hallucinated references in scientific papers then there is some obvious intent to deceive there