I had the Dodge 2.2 Turbo I engine with a 3000 RPM idle problem. The AIS controls the idle speed by letting air bypass the throttle plate. All it is is a valve that opens up to let additional air into the throttle body. It has a semi-circular opening, with a semi-circular shutter that opens or closes to allow air to bypass the throttle plate.
The shutter is designed to rotate 180 degrees, and then it hits a stop to keep it from rotating any further. The shutter sits inside a circular sleeve, and the sleeve is what's turned by the motor. The sleeve holds the shutter snugly enough to turn it when it's free to move, but the sleeve can continue to move when the shutter has stopped, so the motor won't burn out accidently.
Not only is the shutter designed to rotate, it is spring loaded, so it can be pressed down into the motor housing. Doing so makes a closed shutter effectively useless, because air can bypass it. I believe the spring is there to allow air bypass even if the shutter is closed, if there is enough of a pressure differential in the throttle body to force the shutter down into the housing and let air bypass it, perhaps to avoid other damage.
There is one flaw with this design -- when the shutter is depressed, it can be moved past its stops. If that happens, and the shutter then pops back up, it gets locked 180 degrees out of phase. When the computer tries to close it, it is actually opening it. It will keep trying to close it until it opens full up, and then your idle is racing at 3000 RPMs. The computer keeps trying to close the shutter, but since it's wide open, there's not enough pressure differential to force the shutter back down so it can again go past the stops into its normal position.
This is apparently what happened to mine. For all I could tell, the motor worked fine. Applying voltage to it opened and closed the shutter, only I couldn't tell that it was out of phase, opening when it should close and vice versa.
What compounded my problem was the fact that the shutter was gunked up with carbon and sticking in the sleeve. So it got stuck down and stayed there, allowing it to pass the stops. Eventually, it popped out enough to lock into an out of phase position. That's when the problem wouldn't go away. I finally figured it out when I noticed a rattle in the motor, and that the shutter was moving up and down slightly as I shook it. I was so fed up at that point that I ignored the service manual advice to the contrary, and took apart the AIS. I didn't ruin it -- it's a pretty simple design. I cleaned the parts and got the shutter to where it wasn't binding in the sleeve anymore. I then reassembled it (praying I had it in phase, which I did) and everything was fine (if it still raced, I would have taken it apart and reassembled 180 degrees around).
After "fixing" the problem, I went for a test drive. Everything was OK for a few minutes, but then the racing returned! Back to the garage where I found the following (warning -- more conjecture coming). In the top of the throttle body, there are two more ways for air to bypass the throttle plate. One is an orifice which is controlled by a set screw (idle set screw?). A while back I had apparently closed off the orifice -- some carbon gunking didn't help either. Running the engine at various speeds created the pressure differential needed (which couldn't be compensated for by the closed off orifice), and the AIS shutter passed its stops again. I backed the screw off, re-fixed the AIS, and got the engine running again. I then sprayed some carb cleaner into that orifice. No problems since then!
The second air bypass is a small hole in the side of the throttle opening just above the plate. It had some carbon buildup too, so I sprayed it out as well. (Yes, I know, I should use throttle body cleaner. I just didn't have any handy.)