I'd expect the "receiver"/pot to deform the field enough as to not matter anymore at those distances. Plus, inductive hobs around here tend to detect the presen

I'd expect the "receiver"/pot to deform the field enough as to not matter anymore at those distances. Plus, inductive hobs around here tend to detect the presence of a pot and not just try pumping 2kW of energy into "nothing". And you're comparing a dynamically pulsed EM field with the static M field of the alignment magnets - they might be able to heat the reed, but might just not be able actually trigger it (edit: to clarify, that's my guess; I'd expect manufacturers of either devices [ICDs and hobs] to test for potentially lethal problems).