Consider something it may be easier to think about:
A car is accelerating forwards.
Therefore there must be a net, unbalanced, force acting forwards.
The car is driven from the motor acting through the wheels in contact with the road surface ... so the force accelerating the car must be the contact force with the road ... which is friction. Therefore, the dominant friction-force on the accelerating car must be pointing forwards (in the direction of acceleration).
Therefore - friction does not always oppose motion. What it does is sap energy ... which is why the car needs a motor.
Now... bearing this in mind, instead of thinking "friction" think "net force", and do the problems:
1. the net force on the ball must be pointing east if it is accelerating east ... which shows you that the question is under-specified: if it was accelerating and now moves at a constant speed, then there must be another net friction force opposing the motion.
2. if the bike is being pedalled, it is being driven by the back wheel - accelerating the bike forward - so the net force from the back wheel must point forward. The front wheel will resist this motion, contributing to the drag, but not as much since the whole bike is accelerating. The context smooths out questions that cause trouble with #1.
3. when you stop pedalling a bike, it slows down (accelerating in the opposite direction)... you've stopped driving the back wheel so its net friction changes direction.
3 does not contradict 1 because the situation is different... made clear if we consider the case for 1 where the ball is accelerating east and not just at a constant speed. But still, it is not decelerating.
I'd complain about #1 ... the nature of the motion is not specified and could be accelerating, decelerating, or uniform. The mere fact of motion in a particular direction is not enough to tell the direction of the net force. To properly answer these questions we need to use metadata - information not explicitly supplied in the question. IRL: this is almost always the case.