I'm not sure how the design prevents air flowing over the winglets while leaned over in a flat corner doesn't generate some horizontal force.
Likewise.
If adjustable winglets are permitted, I can imagine a way they would work. Angle them downwards at 45 degees, so that when banked over, the outer one is roughly horizontal, the inner winglet is vertical.
An on-board sensor detects that the bike is cornering, so adjusts the inner winglet so that it develops lift instead of downforce. and pulls the bike towards the centre of the turn. The outer (upper) winglet is now roughly horizontal, so develops downforce, helping to increase adhesion. During straight line running the winglets could be adjusted for minimum drag unless the front end is beginning to lift, in which case both winglets get adjusted for downforce.