First, microwave ovens are not tuned to water. They're tuned to an ISM (industrial, scientific, medical) band that the FCC allows for high power use:
https://www.wtamu.edu/~cbaird/sq/20...icrowaves-in-a-microwave-oven-tuned-to-water/
Commercial microwave driers are often tuned to other frequencies:
https://www.sairem.com/microwave-ra...rying/microwave-rf-drying-industrial-process/
Which are also typically ISM bands:
https://en.wikipedia.org/wiki/ISM_band
Microwave ovens work by dielectric heating. Water is an electrically polarized molecule that will align itself with an electric field. If that field is inverted, the molecule tries to flip, and then again, and then again. That vibration heats the molecules. It doesn't matter what the frequency is, it's the back and forth motion that creates the heat. Other polarized molecules will heat too, but water is present in most food so it tends to be what heats the food.
So saying that tissue being water based means that it dissipates heat is exactly backwards. It's the water that's being heated.
Water questions aside, heat dissipation is a critical factor. The limits are set with specific tissues-- the eyes and testes-- in mind because those have the least blood flow and thus dissipate the least heat (not because of the water, but because of the lack of circulation):
https://www.fcc.gov/engineering-tec...ision/radio-frequency-safety/faq/rf-safety#Q5
We all seem to agree that holding an open microwave oven against our heads is bad. So there is some safety limit. The question then is where should that limit be? Less than 1000W is a start.
The FCC SAR limit is 1.6W/kg for us common folk:
https://transition.fcc.gov/bureaus/oet/info/documents/bulletins/oet65/oet65c.pdf
That limit is based on finding in an ANSI guideline that suggested a limit of 4W/kg as a safety margin for the general population but accepting 20W/kg for "controlled environments" (basically professionals exposed as part of their work):
http://emfguide.itu.int/pdfs/C95.1-2005.pdf
This is why I'm not too worried about the levels shown in the newspaper report-- the FCC numbers are conservative enough that minor violations probably aren't an issue. The FCC numbers are also averaged over 30min, which isn't my normal use case. Still, shipping a product that isn't compliant with the law should be taken seriously and saying I'm not very worried about iPhones is very different from saying RF is harmless.
The FCC specifies their limits as averaged over 1g of tissue because they're not really concerned with your phone heating up your whole body, but are worried about it heating up certain very sensitive tissues.
Finally, there seems to be the constant confusion of near fields and far fields. When the phone is pressed to the ear, the far fields are present, but there's also near fields which can potentially be much more intense:
https://ocw.mit.edu/courses/electri...s-spring-2009/readings/MIT6_013S09_chap10.pdf
The far field gets stronger as you get closer to the source as a function of 1/(r^2) (essentially the product of eqs 10.2.8 and 10.2.9 in the above reference). The near field gets stronger as 1/(r^5) (essentially the product of 10.2.15 and 10.2.16). Far from the antenna, the near field barely registers, but close to the antenna it dominates.
Near field shapes are also notoriously difficult to predict and model and they're very sensitive to materials close to the antenna which can lead to hot-spots.
The near field has another interesting characteristic described right after 10.2.16: it's reactive. Transmitted power from the phone is lost and drains the battery whether there is a receiver listening or not.
Reactive power won't really be measured at the battery terminals-- the energy is recovered by the antenna each cycle.
All of which is to say that arguments about how much power the radio transmits don't tell you much of anything about the SAR. The SAR is driven by the near field and we don't have any datasheet spec that can tell us what the near field strength is.
So quoting transmit power numbers doesn't indicate anything about the safety of the device.
If Apple and Samsung passed FCC compliance before they shipped, I'm not too worried. Maybe they gamed the system to make the results from their golden test unit look better than a typical production unit, but it would be shocking if they pushed it so far that whatever they're actually producing is dangerous. I think it's better than even odds that the newspaper doesn't understand how to test, but if it turns out they're right then there should be a reckoning. My guess is that will take the form of tighter oversight of testing labs by the FCC for a while. I doubt Apple will lose a lawsuit because they probably followed the letter if not the spirit of the regulations.
Again, none of this supports blanket statements like "RF is harmless". Wireless charging is also RF and in that case the scenario can be very different.