This talk of a passenger decoupling the autopilot can be pretty much discounted. They’d have to either press two buttons on the centre console and disengage both channels or reach across and press the flight director buttons on the pilot side or reach across and press the decouple button on the pilot cyclic. All highly unlikely (bordering on ridiculous)
The mention of someone being a front seat passenger and overriding the controls is plausible if they have a cyclic in front of them, as they can interfere with flight controls and grabbing a cyclic in a stressful situation could cause all sorts of problems for the pilot. Its unlikely they could press down on the collective (unlikely because you need to press a trigger under the collective handgrip to move collective. )
Some polarising sunglasses definitely prevent you from viewing the screens in a C++ from certain angles, but I’m not sure about a B model.
All of this talk is supposition. We will likely never know the real cause of this accident, but if nothing else it should make the helicopter pilot community maybe stop and think a second longer when they next take that flight in deteriorating conditions. We can only live in hope.