Can energy be extracted from clouds? Can cloud charge imbalance be used as an energy source?
First off quite some energy must be present in clouds: a lightning path is quite long, and electrical breakdown of air requires about 1MV/m. Most (many, smaller) electric discharges are not seen because they occur within clouds. The electric energy built up can also be lost in other ways (i.e. Lorentz force).
Some of it comes to ground and when it happens it has such large peak power (on average about a trillion watts) that it can not efficiently be harvested.
I assume that charges build up quite continuously during the lifetime of a cloud: if we knew what cloud to select for discharge we could control (by selecting the right cloud or part of a cloud) the power flux, though not easy its hard to rule out as impossible.
Now how exactly do I propose to do this?
Air can be made conducting through bruteforce electrical breakdown (not efficient), or by heating it (not efficient) or by an energy form encouraging with higher specificity the ionization of specific molecules present in air (be it oxygen or nitrogen or ...).
Apparently non-lethal weapons were designed which do not use projectiles:
a UV laser with 2 optical, parallel paths, each connected to an electrode at the weapon side can tetanize (stimulate at such a pulse rate as to keep the muscle contracted) with specificity any of the 3 broad muscle types (voluntary skeletal muscle, involuntary smooth muscle, and cardiac muscle all respond to different pulse patterns/frequencies). A weapon on same principle was designed to stop cars up to 2km by frying the control chip in the car.
According to hstvi (not sure how credible this source is) ionizing wavelengths with lasers can make a path of air conductive to 2km:

"With an ultraviolet wavelength of 193 nanometers, the maximum
  effective range is about 100 meters. The effective range increases to
  two kilometers when a wavelength of 248 nanometers and a more intense
  beam are used."

Now for our clouds we would not need 2 optical paths to the laser.
Problems I envision: clouds heights are on kilometer scales, so possibly stages of such lasers would need to be mounted on weatherballoons, preferably locked by cable to a certain height/location.
Each laser would could be surrounded by a thick piece of transparent insulator (glass?) so as to allow UV to pass but not current to enter the setup. On the glass electrodes are placed which are connected to a similar electrode on say a retroreflector of this balloon (a lower one targets this balloon).
If forming a conductor with the lasers consumes more energy than would be got from the lesser charged clouds (or parts of them), we may need to allow electrical breakdown to occur at which point switching of the laser beam would not stop the lightning.
Problem lightning?
 A: Credibility


*

*Your source is at first not really credible to me as a physicist and how should air be made conductive? This can only be done at very high pressure, i.e. we do not talk about ideal gases as in the atmosphere anymore. The particles which would be ionized still have a very high velocity such that such a path is unrealistic.

*The distance of 2 km by far not long enough but that is a minor point.

*You should calculate at first the laser power consumed for such ionization. You are exactly doing what the atmosphere does by itself. 

A: I do not have an answer, but more questions and ideas, which are probably more valuable. Also as a preface I am being slightly loose with my thoughts, as this is not strictly a scientific discussion to start with. 
The laser conduction path likely works by ionizing and charging the molecules in air. I can see two major sources of loss in this system. One is the energy required to generate the laser. Two is the ohmic I^2*R losses in moving current from the cloud to a system on the ground. As you mentioned this can be as long as a 2 km transmission path. I would bet money that the efficiency (Power Generated divided by Power Input to the system) would be very low, not justifying the effort of the whole thing. Have you thought about these two major sources of power loss? A more specific question is how much power did your non-lethal weapon require to run?
I would propose harvesting the energy in a different way. Imagine the charged molecules of the cloud moving around. In doing so, they create electric and magnetic fields. There is a device which can create current from B-fields moving around its axis... we call it a wire. I cannot justify this, but I imagine there is a passive device that can use these stray B-fields to induce currents. We would just need to float it up into the cloud and keep it attached with a conductor. A battery on the bottom end of the balloon will look like a much lower impedance to GND than other paths, and may be able to charge consistently while the balloon is in the cloud (instead of allowing the potential to build up to a point where it can arc to GND via lightning). 
There may be an additional benefit to any system of this type (regardless of the topology) which has not been considered yet. In harvesting the energy from a cloud, we are taking it from the cloud and moving it somewhere else. I wonder if in doing this process, the reduced potential in the cloud can (1) reduce lightning strikes or (2) have predictable meteorological effects. If (2) is true, then we have simultaneously harvested energy and controlled weather patterns. 
Best Regards,
Brady
A: Of course it can. And if you hadn't listened the the smart "physicists" here but had pondered the possibility instead, then maybe it would be you getting the medal and recognition, and not Samuil Kruglyak:
http://www.day.kiev.ua/en/article/economy/one-kilowatt-003-hryvnias
Sorry, you have missed the boat.
A: Of course you can stick a large metallic pole into a cloud and, charge a battery with any discharge that may or more often may not occur - using really tough electronics. This would be as unreliable an energy supply as a deluge every five years in the dried out desert is to water supply.
