I think the "universal" notion is very interesting and, yes, essential for AI to ever replace the SPG in rec diving. Wookie, I am surprised at the number of failures you see, but, in fairness, a battery dying is not a "failure" of AI, but of the diver not being in control of their gear. Without naming brands, my wife and I dive the same computer. Transmitters are individually coded (in fact, I can see her air supply in a separate window, and she can see mine) and so have no interference with each other or from other transmitters. No interference from the several different strobes I use, nor from any DPV use. Reliability is 100% over about 200 dives (on each computer). No water intrusion like on a hose mounted AI, because the transmitter is securely seated on the first stage. The predictive gas usage algorithm signals ascent so as to get us to the surface with our pre-set reserve with very good accuracy, no matter the dive conditions. It is superior to the SPG in all respects.
So, most of the problems can be (and have been) solved, but it seems that different manufacturers are at different stages of progress.
The only gas monitoring failure I have experienced was when my wife's SPG was slowly getting stuck at higher pressure than what she actually had due to some internal corrosion-- a very insidious and dangerous failure mode. Having the AI alerted her to this and luckily we were at Blue Heron in 18 feet of water. So, an SPG requires caution as well.
And yet, and yet, and yet . . . . I am old school and still have an SPG which I check maybe once during the dive.