Maybe it shouldn't be used for offside unless to prove a clear and obvious error.
Anything else is claiming a level of infallibility that's either impossible or undesirable. The claim has come from someone that it's possible to judge offside by a millimetre. I'm not sure who has said this but that appears to be the claim. It's a bit ridiculous if that's the level of precision they're working to.
Someone correct me here if my reasoning is wrong.
Two players walking in opposite directions at 1 m/s are each moving 1000 mm/s. For a given point on an attacker to move from onside to offside by 1 mm requires each player to move only 0.5 mm. This movement will occur in 0.5 millisecond ... at walking speed. When a player kicks a football, contact between ball and foot has been measured to last between about 7.5 milliseconds and 8.5 milliseconds, depending on the force applied: I believe the smaller the force applied, the longer the time of contact. Clearly, to be accurate according to the laws of football, the offside position of our given point has to be judged at the first half millisecond of that time interval. That, I believe, is the implication of claiming that the system can judge offside to the millimetre. The margin of error is huge, 1600%. And that when two players are moving at walking speed in opposite directions.
https://dspace.lboro.ac.uk/dspace-jspui/bitstream/2134/11470/7/Measurement of Contact Time in Short Duration Sports Ball Impacts.pdf
Please, somebody prove me wrong.