What frequency range exactly are they talking about?
Wondering how they can ever make sure that it would work on a variety of devices. It may work on some specific devices that have enough audio bandwidth all the way to the speakers not to attenuate frequencies above ~20KHz too much and conversely have microphones that are sensitive enough at those frequencies to capture some usable signal, but it would certainly not work in the general case.
Since all modern devices use digital audio, making use of ultrasounds would also imply the standard sampling frequency of 96kHz probably or above (the standard 44.1kHz or 48kHz aren't high enough to get anything really usable above 20kHz, it would be aliased and thus become audible...) As for capturing, I don't know of any phone or tablet that captures sound from their microphones at above 48kHz.
A lot of the internal speakers of TV sets are crap (let alone those in phones and tablets) and unable to emit any amount of usable acoustic energy at ultrasonic frequencies (again there may be unwanted exceptions but you couldn't rely on that), and nowadays a lot, if not most mobile phones microphones are MEMS-based, with rather wildly varying characteristics from model to model. MEMS microphones usually (but not all) have a frequency response usable above 20kHz (but pretty much never documented), but again you'd be playing with just odds and again, there is the sampling rate issue and the analog front-end as well (whether it's embedded in the microphone itself or not).
Would be interested in getting more technical info on this "technology", as it "sounds" fishy at first thought.