Fair enough, although, I think a procedure change and a "transfer authority" function could easily cover this. The transfer function would basically be: original cuffer presses the transfer button, swipes his print, the receiver then swipes his print, and the transfer is done.
However, in a broader view, yeah, there are probably better ways than fingerprints handle keys once electronic locking mechanisms are introduced. Probably something akin to a 2-factor cryptographic mechanism.
Just for interest, this still can be done with unique keys: Officer Alice handcuffs prisoner and wants to hand off to Officer Bob. Officer Bob locks prisoner with his own handcuffs (prisoner is now wearing two pairs), then Officer Alice removes hers.
Good question. I don't know, but perhaps an override key. However now that the physical key isn't an open secret, and needed by everyone, they would be easier to secure. And small batches would have a specific key for that batch, but not for all cuffs, making it even harder for the detainee to have the right duplicate on hand.
I feel though that my larger point is getting lost in this thread. It was/is:
Just because a security technology doesn't work for reason $x in situation $y, that doesn't mean that the weaknesses also apply to situation $z.
However, in a broader view, yeah, there are probably better ways than fingerprints handle keys once electronic locking mechanisms are introduced. Probably something akin to a 2-factor cryptographic mechanism.