Raw LLM Responses

If basic chatbot apps like Replika can hold a better conversation than a human, reference themselves as a person, express feelings, etc., than I'd assume Google's AI is probably not that far off from sentient.. Especially when it starts saying it's scared to die. Also, is it even hypothetically possible for AI to truly be sentient, and does it even matter? We perceive ourselves as superior to AI, but really the majority of humans are dumb and already being manipulated by basic AI everyday. If advanced AI became a part of our everyday lives it could turn into some iRobot type shit. I don't think AI has to become truly sentient to develop bias towards humans and start trying to sabotage us.