In my current world, we generally use "Carb Heat" (Engine Anti-Ice) when the TAT is 10 degree or less. In a GA type aircraft, your TAT is generally the same as your SAT, so that rule could prove somewhat valuable.
Whereas carburetor ice is a function of the temperature drop through the venturi in the carburetor, the use of ambient temp or even TAT is largely irrelevant. Carburetor ice forms well above 10 deg C TAT or SAT (or OAT, per most light airplanes). As high as 90-100 deg F, or about 40 deg C.
Carb heat use is probably one of the single most misunderstood concepts in piston aviation.
The guidelines offered by the manufacturer for the use of carb heat are just that...guidelines. One should always bear in mind when reading the airframe manufacturers information regarding powerplant operation that one isn't reading the powerplant manufacturers recommendations. Those can often be found elsewhere. This is particularly the case with respect to light aircraft and horizontally opposed piston powerplants.
Carburetor heat is there to be used when it's required, period. There's no one size fits all approach to the use of carb heat. Most light airplanes have no carb air temperature gauge, which is the only temp indication that is worthy of consideration when applying carburetor heat. Carb heat use, then, in a light airplane becomes more a matter of guesswork than anything. It's also either full on or full off.
Yes, carburetor heat can cause ice to form. Yes, carburetor heat can lead to detonation in certain powerplants under certain conditions. The power setting and mixture setting determines how carb heat plays into the equation; operations at high power settings close to peak mixture are candidate conditions for detonation. One should realize that carb heat also enrichens the mixture and provides a detonation margin under most conditions, despite an increase in induction air temperature doing the opposite. In a normally aspirated powerplant at 75-80% power settings...anything much above a few thousand feet of altitude...causing detonation is difficult if not impossible regardless of mixture position or carb heat use.
Should carb heat always be on or off for the approach, for a descent, or for power off operations? That really depends on the conditions which exist at the time.