Studies show that natural language models can be a source of academic misconduct, suggest ways to prevent the problem
