Why don't we use our own light production at night (I mean home, buildings, streets,..., lighting) to charge photovoltaic panels instead of wasting it?
Solar panels work with sunlight. The energy per square meter of light from the sun , depending on the geographic area etc is of order of
8 hour summer day, 40 degree latitude 600 Watts per sq. meter
In one hour a photovoltaic cell of one square meter will provide energy of 600 watthours
Take a light bulb of 100 Watthour . To gather all that irradiance one would need to cover all the walls of the room with photovoltaics, which typically are 30% efficient. So one would gather 30 watthour of that "wasted energy" for an enormous cost in photovoltaics. ( disregarding that photovoltaics should be specially developed for low intensity conditions and as pointed out in the comments that a lot of that energy is in the infrared spectrum ).
In analogy will be the economics for other situations, Take a stadium with its large light sources, the power falls as $1/r^2$, where r is the distance between the light source and the panels.
We don't harvest waste artificial light, because it would be ridiculously expensive to do so. The energy in sunlight is, at full sun, $1000W/m^2$. That's way higher than any artificial light in normal circumstances. So it's far more economic to position solar panels to optimise collection of daylight, rather than to capture artificial light at night.
Any energy collection is going to have significantly less than 100% energy efficiency, unless it's low-grade thermal energy you're collecting. And photovoltaics are typically about 20% efficient. Whereas energy efficiency is pretty much 100% energy efficient: when you cut your energy demand by a gigajoule, the amount of energy consumed is going to drop by a gigajoule (even more if your electricity comes from thermal plant). So if there's an economic incentive to harvest waste light, the economic incentive is going to be even higher to reduce the amount of light created in the first place.
As MSalters noted in the comments - if light isn't going in the direction you want it, the cheapest way to harness it is to use a reflector to redirect it to a useful direction. This generally requires very little investment, and will be 60-90% efficient.
Only in the rarest, freakiest cases will it make sense to use PV to turn artificial light back into electricity: if there's a barrier which will let light pass through, but through which you can't run electric cable, and you've got electricity only on one side of the barrier, but need it on the other. Once every few years, in some freaky lab circumstances or extreme hazardous area, this happens.
Maybe when all the stars start to fade this is necessary, unless our eyes have evolved into something else.
Edit: we probably do not do it because the sun is still active and photocells require rare materials (the ones I have heard of) and if you would have photocells everywhere where would you nail that nice painting you made?
Our light collection systems are horrible. A reliable source says that the best commercial solar panel out there work at a whopping 22% efficiency.
The power generated even in a major city like London is very low and also highly angle dependent.
An interesting idea, however, could be to use a mirrored lens to focus the low-light from various angles into a centralised point rather than focusing on maximising surface area and collecting a significant flux of light.
Perhaps this could be a "leech" system for low-power LEDs from nearby sources.
For sure, even having the best photo voltaic panel produced so far, you would not be able to convert back all the energy. You will have always loss. Anyhow, this idea is very similar to what is currently done with hybrid cars, here part of the kinetic energy is converted back to battery energy when the car is slowing down.