Note you can make LEDs dim without PWM via constant current drivers, just less efficient as the duty cycle is 100%, they look much nicer though
anecdote: I work in a factory, we have thousands of light fixtures. The older constant current LED fixtures have lasted a decade, the new PWM driven lights have had a ~1% failure rate.
Yes, adding an inductor will fix the problem, but that costs money. Not enough customers are complaining about the light flicker to justify the higher price.
In those days TV's were big CRT boxes, illuminated by the electron beam which would race across the screen row by row.
A housefly's "refresh rate" on its eyeballs is much higher than that of a human, so the fly wouldn't see an image as we do, but would presumably find it quite relaxing watching the dot of light slowly move across the screen.
Anything more modern than that, even if it has a similar case, will use less power
This Acuity occ sensor/photocell combo draws 4mA @ 24VDC: https://pdmassets.azureedge.net/library/SS/documents/specshe...
However, my mind developed a similar conjecture to your's: what if the common frequencies of PWM occur within a kinetic visual blindspot of insect vision, like human saccades [0] in the novel Blindsight [1], so that insects are not aware of the LED at all?
0. https://www.ncbi.nlm.nih.gov/books/NBK10991/
1. https://www.rifters.com/real/Blindsight.htm#:~:text=%22There...
They've had many generations to evolve alongside all human false light sources. Insect generations are usually pretty short. Perhaps they just find it evolutionary useful to cluster around human sources of light. Presence of human food, liklihood of bumping into other insects of the same species, warmth, etc.
If human sources of light were interfering somehow, they've had hundreds or thousands of generations to breed out that proclivity.