They should but they won't.
As an example, multiengine general aviation aircraft have worse safety records than single engine aircraft. The reason for that in part is that single engine general aviation pilots fly around knowing that if the engine noises stop, they are very likely going down and better have a plan and a place in mind when that happens. Twin engine general aviation pilots upon suffering an engine failure default to the idea that the remaining engine will get them to an airport for a safe landing. However, the validity of that assumption depends on a lot of factors such as the density altitude, wind, turbulence, weight and balance of the aircraft and the skill and currency of the pilot. In some situations even when flown to perfection, some twin engine aircraft can't maintain altitude and the remaining engine just extends the glide to the eventual crash. The pilots involved in these crashes would have often been far better served by feathering the good engine as well as the bad one and just committing to a dead stick landing. They crash because they become complacent with two engines and they lose the judgement and skill necessary to properly manage an engine failure in a twin engine aircraft, and they fail to recognize when they just need to use the remaining altitude to make a controlled crash landing they can walk away from rather than spinning in from low altitude after losing control trying to make it to a runway. I have two friends that are dead after trying to do exactly that.
The point here is that if you develop genuinely accurate O2 sensors and CO2 sensors, it'll just lead to more complacent CCR divers who become dependent on highly reliable sensors that can, and will, still eventually fail. When that happens those divers will be less likely to survive as they will be less likely to have the necessary skills to identify the fault before it kills them, and/or they will be less prepared with less than adequate bailout.
A large part of technical CCR training is learning how to determine when a sensor may be wrong, learning how to verify which, if any, O2 sensors are correct, and learning various bailout modes to make the most efficient use of the resources you have on hand. With those theoretically perfectly reliable sensors, all that emphasis on managing failures effectively goes away and when things get pear shaped that poorly trained diver will die.
CO2 and scrubber breakthrough is the thing I worry about the least. I buy good quality sorb, I store it in good conditions, I use an easy to pack and as reliable as it gets axial scrubber, I take great care in packing my scrubber, and I've put in a lot of time and effort to know what the limits of that scrubber are under the real world conditions I use it in. I also pay very close attention to any water in the loop, in recognition of what it can do to the scrubber and adjust the dive accordingly.
Many CCR divers do the same because that is how they were trained. However, if you add accurate CO 2 monitoring to the loop, would those safe practices change? Probably. Dive after dive of no CO2 indications would lead to complacency, especially then less diligent practices don't result in CO2 in the loop. That works right up until it doesn't and then you've got a diver who has to bail out due to a CO2 issue that he's been warned about - but one that he could have probably avoided with better scrubber related practices.
You also have a new system that can fail, at the same time you have a diver that is now dependent on CO2 monitoring, rather than carefully planning well within the scrubber limits, pre breathing the unit, and periodically self assessing if the scrubber is working properly.
Exactly.
