Haha, came here to post that song... Anyways, I still hear this every now and then (the sound, not the song) with my Ford C-max 2006 sound system and iPhone 12 mini, but nowhere near as often as back in the 90's and 00's. Back then I also had a separate vibrating module I'd put in my cell phone pocket (many pockets available at that time, Nokia 3210 did not vibrate), it would always vibrate before any ringtone (self programmed AirWulf) would start (of course).

I thought communication between the tower and the device was encrypted?

That's not to say it's not using some archaic weak encryption scheme left in for backwards compatability though.

Not really, speakers just reacted to the bursts of power, because the transmission frequency is way too high for us to hear.

GSM uses time division multiplex, which means it shares the 8-timeslot channel with 6 more devices (7 devices + 1 signalling), meaning your phone transmits for 577us, then waits quietly for all the 7 timeslots that other devices use (so, for 4.039 milliseconds - 7x577us), and again transmits for 577us, and again waits. Bursts of power every ~4.6milliseconds make a noise that we can actually hear at a bit over 200Hz.

Nah, the frequencies are stil around there, they're all too high to hear, but GSM used TDM - time division multiplex.

To simplify, up to 7 users (+1 timeslot for signalling) share the same channel/frequency, but each of them is allowed to transmit only 1/8 of the time (even less, due to guard intervals). This basically means, that your phone transmitted for a bit over 0.5 milliseconds (577us), then waited quietly for 7 more intervals like that for other devices to transmit, then your phone transmitted again, and the affected speakers "buzzed" because of these 577ms bursts of power every 4.616 milliseconds (this causes a buzz in the frequency range we can actually hear).