Another case of real world vs. what the experts say.
True be known, when a threaded fastener fails in tension, the failure occurs in the first thread that is engaged in the nut (closest to the unthreaded shank). The last thread, the one being upset or tack welded sees little if any load.
If the statement in the latter sentence is true, a small weld or even an arc strike on the exposed threads will do little to compromise the integrity of the threaded fastener. The heat introduced by the arc strike or tack weld, which is simply meant to foul the threads so the nut cannot back off, is not sufficient to temper the hardened threads where they engage the first few threads. It is the fear the welder will get carried away in an attempt to do a "good job" that is the cause for alarm. Should the welder actually deposit a weld of sufficient length, the temperature developed might be adequate to temper the hardened fastener and reduce the strength of the assembly.
In consideration of the "unknown factor" I would opt to use a chisel or center punch to upset the threads. However, if the company was to run several samples and "train" the welder on the amount of weld required, I suppose a couple of shear or tensile tests would be sufficient to establish whether the technique employed would render the welded assembly "unsafe.”
Best regards - Al