Your Resmed AirSense can't tell if you are awake or asleep. Since the device is designed to be used while sleeping, all of the programming for event detection is based on normal sleep breathing patterns and what goes wrong with normal sleep breathing when a respiratory event (an obstructive apnea, a hypopnea, or a central (clear airway) apnea occurs.
In other words, the machine is programed to score an apnea of some sort if it detects that is little or no air going in and out of your airways for at least 10 seconds. Resmed machines use an algorithm (called FOC) to determine whether the airway appears to be clear (i.e. open) or obstructed (blocked). The FOC algorithm uses small, rapid oscillations in pressure to determine the patency of the airway. If the results of the FOC are not clear, the event can be scored as an "UA"---undetermined/unknown type apnea.
The machine will score a hypopnea if there is a significant (sudden) reduction in airflow that lasts at least 10 seconds. By sudden, I mean that the bit of flow rate that is labeled a hypopnea shows a significant reduction in airflow as compared to the breaths that immediately precede it.
Since OAs, CAs, and Hs are scored by your machine solely based on what the machine is detecting in the wave flow data, they get scored even if you are wide awake and using the machine.
Now, a lot of people might ask---why would my wake breathing patterns contain something that could be mistaken for an OA or a CA or an H by the machine?
The answer is essentially that wake breathing is a lot less regular than sleep breathing is. We may unconsciously hold our breath for a few seconds (even 10 seconds or more) while we're concentrating on something. Or we may take a few deep, cleansing, relaxing breaths and then skip a breath or two because we exhaled a bit too much CO2, and that reduces our urge to "breath". Or after taking some deep breaths, our breathing may become shallow because we've blown off a bit too much CO2 during the deep exhalations. If we hold our breath and at the same time allow our epiglottis to close off our windpipe (like when we swallow), the FOC algorithm might very well conclude the stoppage of breathing looks like an OA and thus the machine scores it like an OA.
Now some people find that when they're awake and using the machine, they get a whole lot of events scored---those events are not real because the person was awake when they happened. And the person's normal wake breathing pattern is just so much more "ragged" than normal sleep breathing is, that the machine scores events since the machine has no idea that you're awake when you are using it. We just ignore the data that is recorded when we're actually awake.