If ice inside the flask moving at constant velocity were to melt then a creature inside the system could tell whether it was moving or not. But motion is not absolute, hence it is not possible that ice would melt.
If the flask were accelerating, then for constituents of the system, this is equivalent to an applied uniform gravitational field. That is, in this case you must ask if ice would melt due to the action of a gravitational field. It wouldn't because gravitational field only causes a hydrostatic pressure gradient (in which pressure increases in the direction opposite to that of acceleration), and increasing vapor pressure on top of the ice inhibits melting (actually the opposite occurs).
All of this holds if you were to move the flask and its constituents as a whole, while not shaking it any way. In other words, any observer for whom the flask is temporarily at rest w.r.t. him/herself should perceive no other motion of the flask or its constituents. If this is not guaranteed, then you will end up doing work on the flask and its constituents and therefore the internal energy of the flask will increase (this does not necessarily mean temperature change; if phase transition is the only process occurring inside the flask then there will be no temperature change).
P.S. If the flask were to decelerate thereby losing its kinetic energy, then it is not necessary that it should appear as a gain internal energy of the flask. Some external agency might be sucking out the kinetic energy of the flask instead.