Well for starters you need to add an audio capture device, there is no microphone on a Pi.
After that is it certainly possibly to stream over the network what the microphone captures, like streaming audio from a file.
Which means the Pi needs to be on a network somehow, a LAN or Internet.
Once this is all done another machine on the same network could aim at the Pi's network address and listen to the stream.
It's hard work to be an Agent
Also, you do not mention latency, i.e. the time span between a sound being captured by the microphone and its rendition by speakers on the other machine.
Considering encoding and streaming requirements, plus network delays/reliability, and decoding at the other side, I would expect a delay that measures in seconds, perhaps tens of seconds. Radio stations buffer a minute or so to allow the local client to ride over short network glitches.