Gender and content bias in Large Language Models: a case study on Google Gemini 2.0 Flash Experimental