Alarm or no alarm you are changing the load on the circuit, and I don't see what value any reading on the meter may have as you are not really measuring current through the load you intend to check. That said there may be some cases where it does work though. If you short an output device you likely will read the true or very close to true output current, but the load you shorted across will see little or no current and respond accordingly, so if that response disrupts production enough you may not want to do so. Kind of similar will happen with an input type device. Either way something will see current that does not correspond to actual conditions and corrective action will be attempted by the controls to make it right or possibly go into an alarm status.
If you leave out the complication of sensor data changes based on upsetting the process, you will read different values to the extent that the transmitter is not a perfect current source.
But if the transmitter is not a perfect current source, then any voltage drop across the series ammeter is going to upset the circuit too.
Not a dramatic an effect, but still there.
Meanwhile, with a parallel meter, if you do not know the resistance of the loop and of the receiver you are shorting out you will not know whether you have gone out of the operating range of the transmitter or not.
I have to agree with you based on that alone!
When you allow for the possibility of a line-powered transmitter, where the input terminals are providing the current, then placing an ammeter in parallel with the input is going to really mess things up.
So in that specific case, where the PLC is providing the loop power, realolman's question
why doesn't it just short the circuit out?
is very much on point!