As the saying somewhat goes - "Stuff Happens".
Intrinsic safety views electrical devices from the ability of the circuits to create a spark under operation or failure. IS guidelines not only look at energy sources (batteries), heat generation (resistors, light bulbs) and maximum voltages, but also the circuits ability to store enough energy in the inductance and capacitances to create an ignition source for flammable vapors/liquids. IS standards are designed to be very conservative (sometimes seemingly ridiculous) - but with the idea to keep "Stuff" from "happening". Within a few feet of a gas pump is a potential hazardous environment by definition.
The main issue of a cell phone or pager is NOT the RF - it is things like the battery, vibrator motor, and maybe that charge pump for the LED, florescent, or incandescent bulb backlight. From a IS standpoint, a good portion of a cell phone is unsafe.
How many of you remember as kids the pretty arcing/sparking you could get from your slot cars in the dark? Remember those Boy-Scout fire prevention examples using batteries, a little steel wool, and some gasoline vapors?
Stuff happens even when previously examined in detail by very experienced people and put into practice over years. How long were 747 aircraft in service before TWA fight 800 went up in a fireball over Long Island Sound? Supposedly the vapors in the 747 center fuel tank never reached an explosive ratio. They still don't exactly know what the ignition source was, but the fuel probes are suspect. Prior to this incident it was believed the low energy of the fuel level probes (a capacitance measurement device) was safe.
I have no doubt that under the proper conditions, a cell phone or pager could easily cause a fire at a gas pump. It might be rare, but it could, and possibly already has happened.