A voice to voice implementation would be quite useful and fun for interactive experiences. I think its good idea to look into for the future of OpenAI and maybe even sound effects for role play or story-telling would further enhance user interaction. Like a dense forest for example and then OpenAI would play forest sounds for immersion and begin to narrate the tale unfolding. I think this would be a good idea and something to add!