As mentioned, bathrooms as of the 2014 code cycle do not require AFCI.
AFCI breakers need a standard to which they work, and they need to work to prevent arcs; the above video shows they do not. Also, and my biggest gripe with AFCI, is that they need to be reliably tested/testable by means other than the mfg test button. That they cannot be independently tested via a meter/tester, or even field tested by creating an arc condition (as in the above video) is very troubling to me.
I'm no programmer, but it seems to me to be an easy thing to program AFCI to trip reliably under arc conditions. Take a faulty circuit with bad insulation, over driven NM staple, etc., between ungrounded and grounded or grounding conductors. Any arcs will be random in nature, not cyclic or constant. Program breaker to ignore any arcs the first second, or, say, 20 cycles, it's energized, or when amperage changes (such as starting a vacuum cleaner); only have over-amperage protection. This should eliminate nuisance trips from vacuum cleaners and the like. After that, random spikes/dips in amperage in vs amperage out, such as a loose conductor would cause, would trip the breaker (again, same one second/one third second delay in trip so turning the vacuum off wont trip it either). All it would need is an internal clock, a voltmeter, and a way to record such spikes/dips, recognize the imbalance, and trip on an actual arc fault... or am I oversimplifying this to an illogical degree?