You're assuming it'd either be too weak or too strong. They would calibrate it until it was "just right", as they do (or try to do) with pretty much all of their ergonomics decisions.
[0] Anecdotally, I own both a MagSafe 2 device and an iPhone 5.
OK, in travels some countries probably still try to read the creditcard magstripe, that may be a problem for a part of population. Otherwise it's just like worrying about the macbook demagnetizing your floppy disks..
My university ID uses a magnetic strip and RFID. However the university disabled RFID usage on all of the printers, doors, etc. on day zero after people found out you could easily spoof them.
Granted, at least my important cards use chips/NFC now.
Saying that this could "never work" is silly.
The objection that actually does make sense is to point out that Lightning is a) reversible and b) has 8 separate data contacts on each side. It's not just a power connector. Can't really duplicate this with MagSafe or anything similar.
With a Magsafe-like cable, it'll disconnect after the tugging happens, and if the pull from the tug is strong enough, it'll cause it to potentially go a greater distance, since there's no cable connected.
http://www.webosnation.com/hands-on-hp-veer-s-ridiculous-hea...
Make it an electromagnet powered by the USB host/power supply, and use the iPhone accelerometer (or more than one) to detect being yanked by the attached cord and disable the magnet to disconnect easily.
(or do what everyone else is and have an inductive charging matt with no connector).