IR is a thermal measurement -- at the temperatures you mention.
@chemisttree said they are not super accurate. He is correct. Normal body temp readings of the forehead range from 35C to 36.5C in normal healthy humans. At rest, in a 20C room fully clothed.
The reason for the variation in "normal" has to with a large number of physiological responses which affect surface blood flow to the forehead. Pulse rate is one. Lower pulse rate, like when you are at rest, reduces the amount of blood flowing under the skin. Cooler reading. Coming in from outside temperatures above 38C raises the reading. Jogging to the airport raises external BT readings, too.
Not because the instruments are inaccurate. Biology is the problem. The problem is not because of the use of improperly maintained equipment. Or really deficient methodology. The most prominent problem is the implementation. Which is political. See the link below.
Methodology means - the prescribed and laboratory tested use Are you following directions that say your device is meant to work as a body temp scanner. Methodology. No mention in your directions means NO, it has not been verified for that use.
Political means- some guy decides to scan 20 people every ten seconds to save money or time, or to seemingly increase inconvenience
The very most accurate quick body temp is usually considered to be via anal thermometers. To say the least, this is not remotely practical with gangs of 200 airplane passengers.
Please do not start nit picking at something you do not seem to understand fully.
Here is an informed discussion, most of the problems are related not to the equipment:
https://irinfo.org/05-01-2016-chandler/