This is already complex technology, prone to failure and exposed to sabotage.
As an old engineer, I've a gut feeling that
Sprinklers that open when heat melts a valve and let water flow by gravity have decent chances to do more good than harm.
But actively controlled hoses would go crazy and make damages more often than they mitigate a fire.
Compare with the automatic anti-stalling piece of software. Not very complicated neither: check few sensors, pull the stick. But it went crazy on two flights recently.
Or compare with the electronic smoke detectors we have in the houses presently. Good that they don't inundate the rooms every time they sound the alarm without a reason.
This gives a sense of what complexity engineers can reliably master: it's very, very little. We need to fail several times, preferably at the beginning of the career on less important projects, to get this modesty.
Automated or remotely controlled fire hoses would also need electricity, cameras and so on. In a degraded and stressful situation like a fire, you typically lose these resources when you need them.
Firefighters do rely on technology, but the resources are brought from a place away from the fire, they are maintained daily, and are used regularly by people who train for it and have close control over the machines. Quite different from a remotely controlled device supposed to idle for thirty years and work when needed.
How usual sabotage is, we can only guess. How far sabotage goes, we have examples.
At the very Notre-Dame de Paris, dozens of statues were beheaded two decades ago, shortly after a similar sabotage happened in an other European country, friend and ally of France. Imagine: there are hundreds of cops, in uniform or not, in and around the cathedral. The group could enter the site, break the heads bang bang bang, and get away unnoticed.
Or Ariane flight 36. The public learnt about a cloth in a propellant pipe, but there was also a leak in an other engine, a fire in a third, pogo oscillations in a fourth. This flight was "doomed", as they say. Few months later, the same happened at a US launcher, despite they probably expected it.
With that in mind, the usual sprinkler has some resilience. Saboteurs need physical access to a limited location. Melting the valves needs some serious means, cutting the unexposed pipes too.
My proposal with the tank on the ground is already less resilient. More locations can be attacked, and they're easier to access. It needs pressure in the tank or pumps and electricity. If controlled from the ground, sabotage is easier. Whether the net balance is still favourable as for passive sprinklers under a pool?
A remotely-controlled jet, with cameras, transmissions, electricity+electronics+software, is quite vulnerable. Easier to hamper before starting a fire, easier to misuse to make damage without a fire.