I have no idea if this applies to the particular installation in this case, but there is an important safety issue with all autopilots that are able to drive the elevator trim.
If something goes wrong with the direct pitch control (either because the pilot or passenger is pushing/pulling the yoke, or because the autopilot servo (or some other subsystem) driving the pitch axis, has failed) then the problem can be masked by the autopilot's elevator trim action.
Then, some seconds or minutes later, when whatever was messing with the pitch is no longer messing with it (the human input has ceased, or a circuit breaker has tripped on the defective pitch axis) the aircraft is suddenly massively out of pitch trim.
At cruise speed, it's very possible for the required yoke correction to be more than a human pilot can manage. It's even possible at takeoff speeds, in some types, if the elevator trim is wound all the way back prior to rotation.
This is why autopilots are supposed to have additional features to prevent a trim runaway condition. As one example, my KFC225 will disconnect if the pitch trim has been running for more than about 15 seconds. In reality this detection is normally done in software, and American avionics software..............is going to be hardly foolproof, especially at the GA level of quality where an autopilot failure is deemed to be acceptable at any time.
Autopilots are the most wonderful things for reducing cockpit workload and thus promoting safety, but it's important that a pilot flying a plane with an autopilot understands fully how the autopilot is supposed to work.