No, that's part of the preparation. But remember that you have around ~16 moves to choose from each time it's your turn. And you have to prepare a response for all the possible choices your opponent can make.

This of course gets unwieldy very fast. 16^5 is already over a million different games.

As long as it's all in memory, that's fair game. Everybody at the high level is using engines for training/preparation.

This is balanced by the fact that the sheer number of possible lines is huge. Too big for any human to "remember" them all.

It's also not enough to memorize "good" moves from an engine perspective. "Good" engine moves and "good" human moves are different things.

An engine might see a move as "good", because it calculated 50 moves down the line and found an advantage. This means that you'd have to play the next 49 moves in a very precise manner to reach this advantage. One way to think about it, is that there are multiple positions that are analyzed as "draw" by en engine, but 9 grandmasters out of 10 will see one side winning. That's because one side would need to play perfectly, while the other side has much higher tolerance for mistakes (engine assume perfect play on both sides).

TLDR: engine is fair game, during preparation. It's a big deal, it transformed high level chess, but it didn't break the game, and is unlikely to do so.

No, using engines to find lines is a common way to practice (at least, that's my understanding. Obviously, I'm not personally a world class chess player.)

No, that is exactly what they do in opening preparation.

No. But the opponent will do that too. The trick is to find a move that is just slightly worse according to the engine, but that gives a complicated position that's hard to figure out behind the board. That you will then learn all the in and outs of, and memorize. And then hope you get it on the board...

They need hundreds such ideas for a match.

That is the strategy of opening preparation. For the most common openings, GMs often have a 20 move sequence memorized. The time limit of chess matters a lot, so the strategy is to break your opponent out of their preparation with a move that is unintuitive or considered untenable, after you have determined the ways in which it is actually better than widely considered.

The ideal at top level is to get to move 20 or so with a slightly better position, essentially all of your time (because you've been playing memorized moves), and your opponent already having used most of their time.

It isn’t, partly because you needed to predict the scenario would come up and partly because your opponent can do the exact same thing.

That would depend on the opponent doing the same, and picking the same engine and settings for their side too.