It also ignores that each successive generation of wireless is lower power than the last, so the will be more antennas, but operating at much lower power, for a lower effective dose.
I'm not saying that means it's dangerous, just trying to clarify.
If you assume that people use their 5G phones to transmit a terabyte per month because the speed is so convenient, while they used their EDGE phones for a kilobyte per month, that'a a higher total load, so you can reasonably say that 5G involves higher radiation. Or you can compare for the same transmission, and find that it's lower.
You may also assume that the damage done isn't linear with the transmission power or operating frequency, which complicates matters even more. But generally, 5G uses less power to do the same work than 4G, which in turn uses less… because in each new generation, there's a little better error correction and 64QAM is replaced by 128QAM or so, and so they reduce transmission power a little until the reduced power is again just enough for the better error correction and denser codec.
IMO, a lot of this comes down to "doing things is more dangerous than not", which is nearly always true. Sports may keep you fit but your risk of breaking a leg is higher than if you were sitting in front of a TV. Transmitting data is riskier than not. But also more useful, because not watching youtube in HD is clearly useless, while watching them is worth some risk.
With that in mind, what's the relative transmission power for 4G vs 5G if both are using all their respective maximum bandwidth?
Or the same if you want to compare early 5G with late 4G, because this is really the same thing. They deploy improvements all the time, and after some number of improvements they rename.