Well, not quite.
The inherent danger in a rebreather is that the gas in the loop is not the same as in the tank(s).
You can still breath from the loop, but you don't know if the gas supports life.
You have to differ between semi-closed units and closed ciruit rebreathers.
Current production civilian SCR use pre-mixed gas, for most part nitrox.
Those are the Drägers, Submatix and Azimuth. Gas is added via constant flow, which is usually higher than the diver's breathing rate. The excess gets vented, hence semi-closed. There are bubbles, but less and smaller than OC. Say you use Nitrox32. In the loop you will usually have a few % less O2 than in the tanks since you use some breathing from it. As an example let's say 30%. But the fraction of O2 is less important than it's partial pressure.
At 30 msw the pO2 for the mix would be 1.28 ata, for the loop 1.2 ata. All is fine, it's below the normal rec limit of 1.45 ata.
If you are at 10 msw, the pO2 at 2 ata would be 0.64 for Nitrox32, but only 0.6 in the loop.
When the diver does a lot of work, he breathes harder and uses more O2. The O2 percentage goes down. So if you're out of shape, swimming against a nasty current, anything like that and the pO2 will drop. Say to 0.25 at the surface.
At 30 msw it will now be 1.0 ata, still good to go. At 10 msw it'll only be 0.5 ata and that's as low as you want to get. Now ascent and you may have problems. If for some reason it falls below 0.16 ata you'll likely loose conciousness.
Usual protocol for SCR is to flush the loop (exhale it into the water and get a fresh refill from the tank). A pO2 monitor obviously can help here, and a single cell monitor or deco computer with an integrated cell is enough.
There are other SCR systems, such as the Halcyon, that are keyed to the RMV of the diver. They expel a percentage of the gas on every breath you take and replace it with fresh gas. The pO2 is thus more stable and gas usage more efficient, but they generally also a bit more complicated and expensive.
For CCR, where two different gases are used and the loop content is mixed on the fly to maintain a chosen setpoint three O2 sensors are normal.
In addition there can be problems with gas supply, valves, leaks etc, all influencing or preventing gas addition. While shallow depth may have played a role in this accident, high excertion is unlikely. But trouble with the gas addition would explain the diver passing out and drowning. Even more so if maintainance or modifications were done improperly.
I doubt CO2 was the problem unless there was either no or spent sorb in the cannister.
The Azi's radial scrubber is one of the best designs currently available.