"Consipiracy theorists" have said the government has had this technology for a while. I've considered this particular one to be very likely to be true, because even before this story, it was just so obvious it should be feasible with basic science. I could think of a couple of basic ways to beam "something audible" (not necessarily "sound" but something that would be audible to the system as a whole) into someone's head even with just my understanding of science.
One of the ones I've considered is that we already have the ultrasound beams that others are discussing, publicly available for years, to the point I've heard them at the supermarket [1]. I suspect with just a slight tweak of the math, you could make it so that with two emitters you could create audible sound that exists only at the intersection of two beams, for instance. And that's just one idea of how this could be done; by no means am I claiming it's all of them. Even if I'm wrong about this one, something will work. Sound is so basic and simple that it is logical it can be produced by all sorts of things, because we already know from experience it is produced by all sorts of things.
This is initial research from a public team. I see no evidence from the story that they used any tools that haven't been available for 40-50+ years, albeit not as cheaply or as compactly, of course. (i.e., in contrast to the ultrasonic beam which actually requires non-trivial computational power to attain) I don't find it very hard to believe other non-public entities have the refined-over-decades version of this idea.
Combine this with a bit of chemistry and yeah, you could probably play some serious mind games with people. From what I see, it would be easy, like, give the Mythbusters a couple of weeks and take all their scruples away and they could probably bash out a "plausible" prototype of the process from where they are (not terribly expert in the relevant fields). Even just publicly-known psychoactive drugs provide a lot of tools to play with.
I am not claiming that if this technology does exist that it is being used for any particular purpose. I'm just saying with all that black budget and the obvious utility to intelligence agencies of such basic technology it's very easy to believe that they have it, and that they have for a while, because it isn't some wild hypothesis about unknown science or new forces or aliens or any sort of thing they're suppressing from public knowledge, but rather a fairly boringly-straightforward application of relatively basic science, as shown by this exact article. An infrared laser and some modest cleverness in what you do with it. Other comments speculate about putting together a DIY rig to do this. Not complicated.
[1]: They experimented with playing ads to people waiting in line to checkout. It must have been deeply unpopular, it didn't last long. As a nerd I enjoyed the chance the first couple of times to hear it and play with the parameters and see what happens if I bounced it off my cell phone, how sharp the boundary was, and such, but it would have gotten old pretty fast after that, and most people wouldn't even be that entertained.
Edit: Another example of the sort of thing that I think could lead to sound production with some tweaks, here's a video from 2014 of a mid-air laser system generating enough plasma in the air to be visible: https://www.youtube.com/watch?v=GNoOiXkXmYQ You can hear it makes a very characteristic squealing sound. Do something less violent (much, much less violent) to someone's body, and what are the odds that you could get sound out of it with some years of development? There's just so many avenues to create sound, one of them is going to work.