It is certainly possible to have fully mechanical helicopters without computer stabilisation, as we had them before we had computers. But increasingly, we're seeing more and more 'fly by wire' where any linkage between pilot (or indeed, car driver) and the resultant reaction of the vehicle, is electronic. Indeed, it just cost my brother a large wedge of cash because a "throttle sensor" went, and it turns out, throttle (on his distinctly non-fancy, middle of the range, normal car) is computer controlled, not the old-fashioned cable and a bit of metal linkage it used to be, when you could replace a cable for £1.50. Now, it's a £500 job requiring the entire pedal assemby because you csn't buy the stupid sensor separately.
Sorry, got sidetracked.
But on point, most modern airliners are fly by wire, and they deal with the "if the system fails" issue with multi-level redundancies.
So ..... flight assist? Why not? After all, if you have computerised and/or electrical sensor/motor control systems, if they and all the backup fail, you're no worse off than if a cable broke in the old days. Either way, you're probably coming down .... inelegantly.
There's a direct analogy with learning to fly model heli's. Some people say to learn on twin-rotor coaxials, because they're far easier, being vastly more stable.
Others, including me, say that they don't teach you to fly 'proper' helis because you don't develop the feel for a highly unstable craft, and when you do step up, you have to unlearn bad habits. Rather, learn in a highly unstable collective pitch model BUT with a degree of stabilisation, and with both range limits and control sensitivites turned well down. After you've got the basic feel, turn the stabilisation off, and prepare for some crashes. After you've got that, start increasing control sensitivites (I.e. the amount the aircraft responds to a given amount of stick movement). Once you've got that, reduce then remove range limits and you've got full control.
You can do that with model helis because of motorised control linkages, computerised radios and on-board stabilisation software. However, if you want full acrobatic capabilities, you have to have sensitive controls and the AI turned off.
What I've never seen, if they exist, is a quad with manual linkages, etc. Is it even possible to fly a quad without at least basic AI to turn stick movements into respective motor reactions? Could a human react fast enough to directly control all four motors?
If not, then there's always a degree of computer control and the only issue is the balance between what it HAS to do, and the degree to which we override it.
I can fly a drone and get it to do basically what I want. But ever seen the Drone Racing League? Current season just showing on TV now. Can I fly a (FPV) drone like that? Hell, no. I mean, I can drive a car on the roads, but it doesn't make me capable of driving an F1 car.
Anyone interested in drones, check out what those DRL guys can do. It's eye-opening, and rather jaw-dropping just how fast those things are. The world final was at Alexandra Palace (London) recently (just checked, 13tb June), and is showing (IIRC) next week, with the Berlin semi's I think last night.